diff --git "a/test/music-avqa-test_mm_video+audio.json" "b/test/music-avqa-test_mm_video+audio.json"
new file mode 100644--- /dev/null
+++ "b/test/music-avqa-test_mm_video+audio.json"
@@ -0,0 +1,191711 @@
+[
+    {
+        "id": 12,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000093.wav"
+            ]
+        }
+    },
+    {
+        "id": 13,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000153.wav"
+            ]
+        }
+    },
+    {
+        "id": 28,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000086.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000086.wav"
+            ]
+        }
+    },
+    {
+        "id": 35,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000100.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000100.wav"
+            ]
+        }
+    },
+    {
+        "id": 48,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000023.wav"
+            ]
+        }
+    },
+    {
+        "id": 55,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000224.wav"
+            ]
+        }
+    },
+    {
+        "id": 66,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000217.wav"
+            ]
+        }
+    },
+    {
+        "id": 73,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000128.wav"
+            ]
+        }
+    },
+    {
+        "id": 82,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000153.wav"
+            ]
+        }
+    },
+    {
+        "id": 83,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000048.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000048.wav"
+            ]
+        }
+    },
+    {
+        "id": 85,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000226.wav"
+            ]
+        }
+    },
+    {
+        "id": 89,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000113.wav"
+            ]
+        }
+    },
+    {
+        "id": 90,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000017.wav"
+            ]
+        }
+    },
+    {
+        "id": 93,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000151.wav"
+            ]
+        }
+    },
+    {
+        "id": 95,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000045.wav"
+            ]
+        }
+    },
+    {
+        "id": 96,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000213.wav"
+            ]
+        }
+    },
+    {
+        "id": 103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000291.wav"
+            ]
+        }
+    },
+    {
+        "id": 108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000172.wav"
+            ]
+        }
+    },
+    {
+        "id": 111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000201.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000201.wav"
+            ]
+        }
+    },
+    {
+        "id": 112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000048.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000048.wav"
+            ]
+        }
+    },
+    {
+        "id": 116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000191.wav"
+            ]
+        }
+    },
+    {
+        "id": 118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000030.wav"
+            ]
+        }
+    },
+    {
+        "id": 122,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000141.wav"
+            ]
+        }
+    },
+    {
+        "id": 125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000249.wav"
+            ]
+        }
+    },
+    {
+        "id": 127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000277.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000277.wav"
+            ]
+        }
+    },
+    {
+        "id": 136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000176.wav"
+            ]
+        }
+    },
+    {
+        "id": 139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000111.wav"
+            ]
+        }
+    },
+    {
+        "id": 146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000071.wav"
+            ]
+        }
+    },
+    {
+        "id": 155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000146.wav"
+            ]
+        }
+    },
+    {
+        "id": 156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000082.wav"
+            ]
+        }
+    },
+    {
+        "id": 160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000272.wav"
+            ]
+        }
+    },
+    {
+        "id": 162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000224.wav"
+            ]
+        }
+    },
+    {
+        "id": 169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000141.wav"
+            ]
+        }
+    },
+    {
+        "id": 174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000012.wav"
+            ]
+        }
+    },
+    {
+        "id": 184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000206.wav"
+            ]
+        }
+    },
+    {
+        "id": 188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000195.wav"
+            ]
+        }
+    },
+    {
+        "id": 190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000189.wav"
+            ]
+        }
+    },
+    {
+        "id": 191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000026.wav"
+            ]
+        }
+    },
+    {
+        "id": 198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000199.wav"
+            ]
+        }
+    },
+    {
+        "id": 200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000296.wav"
+            ]
+        }
+    },
+    {
+        "id": 201,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding electric_bass in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000073.wav"
+            ]
+        }
+    },
+    {
+        "id": 210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000036.wav"
+            ]
+        }
+    },
+    {
+        "id": 213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000285.wav"
+            ]
+        }
+    },
+    {
+        "id": 224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000368.wav"
+            ]
+        }
+    },
+    {
+        "id": 225,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007197.wav"
+            ]
+        }
+    },
+    {
+        "id": 227,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004394.wav"
+            ]
+        }
+    },
+    {
+        "id": 235,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000034.wav"
+            ]
+        }
+    },
+    {
+        "id": 239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002565.wav"
+            ]
+        }
+    },
+    {
+        "id": 240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004169.wav"
+            ]
+        }
+    },
+    {
+        "id": 254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003311.wav"
+            ]
+        }
+    },
+    {
+        "id": 260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002391.wav"
+            ]
+        }
+    },
+    {
+        "id": 265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002936.wav"
+            ]
+        }
+    },
+    {
+        "id": 267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004539.wav"
+            ]
+        }
+    },
+    {
+        "id": 275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006428.wav"
+            ]
+        }
+    },
+    {
+        "id": 277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005593.wav"
+            ]
+        }
+    },
+    {
+        "id": 281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008185.wav"
+            ]
+        }
+    },
+    {
+        "id": 286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007086.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007086.wav"
+            ]
+        }
+    },
+    {
+        "id": 288,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002986.wav"
+            ]
+        }
+    },
+    {
+        "id": 290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000563.wav"
+            ]
+        }
+    },
+    {
+        "id": 295,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006543.wav"
+            ]
+        }
+    },
+    {
+        "id": 302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007924.wav"
+            ]
+        }
+    },
+    {
+        "id": 312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008385.wav"
+            ]
+        }
+    },
+    {
+        "id": 313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006334.wav"
+            ]
+        }
+    },
+    {
+        "id": 318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002236.wav"
+            ]
+        }
+    },
+    {
+        "id": 323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008170.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008170.wav"
+            ]
+        }
+    },
+    {
+        "id": 324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008174.wav"
+            ]
+        }
+    },
+    {
+        "id": 326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002572.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002572.wav"
+            ]
+        }
+    },
+    {
+        "id": 329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003590.wav"
+            ]
+        }
+    },
+    {
+        "id": 338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005064.wav"
+            ]
+        }
+    },
+    {
+        "id": 339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006608.wav"
+            ]
+        }
+    },
+    {
+        "id": 341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005615.wav"
+            ]
+        }
+    },
+    {
+        "id": 347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002089.wav"
+            ]
+        }
+    },
+    {
+        "id": 350,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004319.wav"
+            ]
+        }
+    },
+    {
+        "id": 370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003963.wav"
+            ]
+        }
+    },
+    {
+        "id": 375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004688.wav"
+            ]
+        }
+    },
+    {
+        "id": 383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000900.wav"
+            ]
+        }
+    },
+    {
+        "id": 385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002165.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002165.wav"
+            ]
+        }
+    },
+    {
+        "id": 387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006715.wav"
+            ]
+        }
+    },
+    {
+        "id": 391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002514.wav"
+            ]
+        }
+    },
+    {
+        "id": 395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001638.wav"
+            ]
+        }
+    },
+    {
+        "id": 398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007636.wav"
+            ]
+        }
+    },
+    {
+        "id": 400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007982.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007982.wav"
+            ]
+        }
+    },
+    {
+        "id": 409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006646.wav"
+            ]
+        }
+    },
+    {
+        "id": 412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007871.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007871.wav"
+            ]
+        }
+    },
+    {
+        "id": 413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008779.wav"
+            ]
+        }
+    },
+    {
+        "id": 417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000726.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000726.wav"
+            ]
+        }
+    },
+    {
+        "id": 422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006991.wav"
+            ]
+        }
+    },
+    {
+        "id": 431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007716.wav"
+            ]
+        }
+    },
+    {
+        "id": 434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008424.wav"
+            ]
+        }
+    },
+    {
+        "id": 441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001842.wav"
+            ]
+        }
+    },
+    {
+        "id": 464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000480.wav"
+            ]
+        }
+    },
+    {
+        "id": 468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003688.wav"
+            ]
+        }
+    },
+    {
+        "id": 474,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002705.wav"
+            ]
+        }
+    },
+    {
+        "id": 496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001919.wav"
+            ]
+        }
+    },
+    {
+        "id": 497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007128.wav"
+            ]
+        }
+    },
+    {
+        "id": 498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002033.wav"
+            ]
+        }
+    },
+    {
+        "id": 499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007490.wav"
+            ]
+        }
+    },
+    {
+        "id": 500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008333.wav"
+            ]
+        }
+    },
+    {
+        "id": 517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004691.wav"
+            ]
+        }
+    },
+    {
+        "id": 523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000807.wav"
+            ]
+        }
+    },
+    {
+        "id": 528,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002998.wav"
+            ]
+        }
+    },
+    {
+        "id": 535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004045.wav"
+            ]
+        }
+    },
+    {
+        "id": 537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000252.wav"
+            ]
+        }
+    },
+    {
+        "id": 543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001672.wav"
+            ]
+        }
+    },
+    {
+        "id": 550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005868.wav"
+            ]
+        }
+    },
+    {
+        "id": 554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001358.wav"
+            ]
+        }
+    },
+    {
+        "id": 558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001927.wav"
+            ]
+        }
+    },
+    {
+        "id": 560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003723.wav"
+            ]
+        }
+    },
+    {
+        "id": 561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000747.wav"
+            ]
+        }
+    },
+    {
+        "id": 564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008193.wav"
+            ]
+        }
+    },
+    {
+        "id": 567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000181.wav"
+            ]
+        }
+    },
+    {
+        "id": 571,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005321.wav"
+            ]
+        }
+    },
+    {
+        "id": 572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008601.wav"
+            ]
+        }
+    },
+    {
+        "id": 578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006971.wav"
+            ]
+        }
+    },
+    {
+        "id": 579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001690.wav"
+            ]
+        }
+    },
+    {
+        "id": 581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000307.wav"
+            ]
+        }
+    },
+    {
+        "id": 583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002156.wav"
+            ]
+        }
+    },
+    {
+        "id": 590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005269.wav"
+            ]
+        }
+    },
+    {
+        "id": 596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006520.wav"
+            ]
+        }
+    },
+    {
+        "id": 597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008486.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008486.wav"
+            ]
+        }
+    },
+    {
+        "id": 601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001405.wav"
+            ]
+        }
+    },
+    {
+        "id": 608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001260.wav"
+            ]
+        }
+    },
+    {
+        "id": 609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008035.wav"
+            ]
+        }
+    },
+    {
+        "id": 615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005990.wav"
+            ]
+        }
+    },
+    {
+        "id": 618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001284.wav"
+            ]
+        }
+    },
+    {
+        "id": 635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005539.wav"
+            ]
+        }
+    },
+    {
+        "id": 637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004794.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004794.wav"
+            ]
+        }
+    },
+    {
+        "id": 640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000461.wav"
+            ]
+        }
+    },
+    {
+        "id": 642,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004146.wav"
+            ]
+        }
+    },
+    {
+        "id": 644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002078.wav"
+            ]
+        }
+    },
+    {
+        "id": 647,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005377.wav"
+            ]
+        }
+    },
+    {
+        "id": 650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007897.wav"
+            ]
+        }
+    },
+    {
+        "id": 653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005406.wav"
+            ]
+        }
+    },
+    {
+        "id": 660,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001484.wav"
+            ]
+        }
+    },
+    {
+        "id": 666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007116.wav"
+            ]
+        }
+    },
+    {
+        "id": 672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007609.wav"
+            ]
+        }
+    },
+    {
+        "id": 673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007315.wav"
+            ]
+        }
+    },
+    {
+        "id": 693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001466.wav"
+            ]
+        }
+    },
+    {
+        "id": 701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002248.wav"
+            ]
+        }
+    },
+    {
+        "id": 714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002349.wav"
+            ]
+        }
+    },
+    {
+        "id": 717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005045.wav"
+            ]
+        }
+    },
+    {
+        "id": 718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001450.wav"
+            ]
+        }
+    },
+    {
+        "id": 723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000993.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000993.wav"
+            ]
+        }
+    },
+    {
+        "id": 737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008616.wav"
+            ]
+        }
+    },
+    {
+        "id": 744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004846.wav"
+            ]
+        }
+    },
+    {
+        "id": 753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006022.wav"
+            ]
+        }
+    },
+    {
+        "id": 755,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002696.wav"
+            ]
+        }
+    },
+    {
+        "id": 759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002747.wav"
+            ]
+        }
+    },
+    {
+        "id": 761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007671.wav"
+            ]
+        }
+    },
+    {
+        "id": 764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005420.wav"
+            ]
+        }
+    },
+    {
+        "id": 766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000326.wav"
+            ]
+        }
+    },
+    {
+        "id": 767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004016.wav"
+            ]
+        }
+    },
+    {
+        "id": 773,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006346.wav"
+            ]
+        }
+    },
+    {
+        "id": 774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008510.wav"
+            ]
+        }
+    },
+    {
+        "id": 782,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002152.wav"
+            ]
+        }
+    },
+    {
+        "id": 783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003577.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003577.wav"
+            ]
+        }
+    },
+    {
+        "id": 786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004315.wav"
+            ]
+        }
+    },
+    {
+        "id": 789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding electric_bass in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002027.wav"
+            ]
+        }
+    },
+    {
+        "id": 790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002796.wav"
+            ]
+        }
+    },
+    {
+        "id": 791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006653.wav"
+            ]
+        }
+    },
+    {
+        "id": 798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004562.wav"
+            ]
+        }
+    },
+    {
+        "id": 809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002480.wav"
+            ]
+        }
+    },
+    {
+        "id": 810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005697.wav"
+            ]
+        }
+    },
+    {
+        "id": 813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002761.wav"
+            ]
+        }
+    },
+    {
+        "id": 815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004794.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004794.wav"
+            ]
+        }
+    },
+    {
+        "id": 818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007468.wav"
+            ]
+        }
+    },
+    {
+        "id": 829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002430.wav"
+            ]
+        }
+    },
+    {
+        "id": 843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006665.wav"
+            ]
+        }
+    },
+    {
+        "id": 849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008032.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008032.wav"
+            ]
+        }
+    },
+    {
+        "id": 851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006913.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006913.wav"
+            ]
+        }
+    },
+    {
+        "id": 853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006359.wav"
+            ]
+        }
+    },
+    {
+        "id": 856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008979.wav"
+            ]
+        }
+    },
+    {
+        "id": 857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004217.wav"
+            ]
+        }
+    },
+    {
+        "id": 865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007770.wav"
+            ]
+        }
+    },
+    {
+        "id": 870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003420.wav"
+            ]
+        }
+    },
+    {
+        "id": 884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003582.wav"
+            ]
+        }
+    },
+    {
+        "id": 885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002679.wav"
+            ]
+        }
+    },
+    {
+        "id": 887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005152.wav"
+            ]
+        }
+    },
+    {
+        "id": 891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005334.wav"
+            ]
+        }
+    },
+    {
+        "id": 901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004155.wav"
+            ]
+        }
+    },
+    {
+        "id": 906,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002865.wav"
+            ]
+        }
+    },
+    {
+        "id": 911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007212.wav"
+            ]
+        }
+    },
+    {
+        "id": 915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007298.wav"
+            ]
+        }
+    },
+    {
+        "id": 923,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004477.wav"
+            ]
+        }
+    },
+    {
+        "id": 927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003878.wav"
+            ]
+        }
+    },
+    {
+        "id": 937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004003.wav"
+            ]
+        }
+    },
+    {
+        "id": 939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000857.wav"
+            ]
+        }
+    },
+    {
+        "id": 942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002212.wav"
+            ]
+        }
+    },
+    {
+        "id": 943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000710.wav"
+            ]
+        }
+    },
+    {
+        "id": 950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004731.wav"
+            ]
+        }
+    },
+    {
+        "id": 951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007853.wav"
+            ]
+        }
+    },
+    {
+        "id": 963,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004938.wav"
+            ]
+        }
+    },
+    {
+        "id": 966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008797.wav"
+            ]
+        }
+    },
+    {
+        "id": 968,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004233.wav"
+            ]
+        }
+    },
+    {
+        "id": 977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001102.wav"
+            ]
+        }
+    },
+    {
+        "id": 981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006264.wav"
+            ]
+        }
+    },
+    {
+        "id": 983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005776.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005776.wav"
+            ]
+        }
+    },
+    {
+        "id": 984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001063.wav"
+            ]
+        }
+    },
+    {
+        "id": 986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004065.wav"
+            ]
+        }
+    },
+    {
+        "id": 990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004519.wav"
+            ]
+        }
+    },
+    {
+        "id": 1000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007834.wav"
+            ]
+        }
+    },
+    {
+        "id": 1017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003971.wav"
+            ]
+        }
+    },
+    {
+        "id": 1022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004648.wav"
+            ]
+        }
+    },
+    {
+        "id": 1025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005240.wav"
+            ]
+        }
+    },
+    {
+        "id": 1026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001440.wav"
+            ]
+        }
+    },
+    {
+        "id": 1036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001581.wav"
+            ]
+        }
+    },
+    {
+        "id": 1038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008339.wav"
+            ]
+        }
+    },
+    {
+        "id": 1043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007242.wav"
+            ]
+        }
+    },
+    {
+        "id": 1046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007227.wav"
+            ]
+        }
+    },
+    {
+        "id": 1049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006299.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006299.wav"
+            ]
+        }
+    },
+    {
+        "id": 1052,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008131.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008131.wav"
+            ]
+        }
+    },
+    {
+        "id": 1057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008667.wav"
+            ]
+        }
+    },
+    {
+        "id": 1067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008392.wav"
+            ]
+        }
+    },
+    {
+        "id": 1069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000936.wav"
+            ]
+        }
+    },
+    {
+        "id": 1074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000311.wav"
+            ]
+        }
+    },
+    {
+        "id": 1079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007938.wav"
+            ]
+        }
+    },
+    {
+        "id": 1090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001641.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001641.wav"
+            ]
+        }
+    },
+    {
+        "id": 1092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005886.wav"
+            ]
+        }
+    },
+    {
+        "id": 1096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003212.wav"
+            ]
+        }
+    },
+    {
+        "id": 1099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006375.wav"
+            ]
+        }
+    },
+    {
+        "id": 1103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001177.wav"
+            ]
+        }
+    },
+    {
+        "id": 1113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003432.wav"
+            ]
+        }
+    },
+    {
+        "id": 1117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008274.wav"
+            ]
+        }
+    },
+    {
+        "id": 1119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000529.wav"
+            ]
+        }
+    },
+    {
+        "id": 1120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002325.wav"
+            ]
+        }
+    },
+    {
+        "id": 1127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002293.wav"
+            ]
+        }
+    },
+    {
+        "id": 1128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000768.wav"
+            ]
+        }
+    },
+    {
+        "id": 1130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001646.wav"
+            ]
+        }
+    },
+    {
+        "id": 1132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000871.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000871.wav"
+            ]
+        }
+    },
+    {
+        "id": 1138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005778.wav"
+            ]
+        }
+    },
+    {
+        "id": 1150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006142.wav"
+            ]
+        }
+    },
+    {
+        "id": 1153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000133.wav"
+            ]
+        }
+    },
+    {
+        "id": 1154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007733.wav"
+            ]
+        }
+    },
+    {
+        "id": 1155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003540.wav"
+            ]
+        }
+    },
+    {
+        "id": 1160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006583.wav"
+            ]
+        }
+    },
+    {
+        "id": 1163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005571.wav"
+            ]
+        }
+    },
+    {
+        "id": 1164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003068.wav"
+            ]
+        }
+    },
+    {
+        "id": 1175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003641.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003641.wav"
+            ]
+        }
+    },
+    {
+        "id": 1177,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006964.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006964.wav"
+            ]
+        }
+    },
+    {
+        "id": 1182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005968.wav"
+            ]
+        }
+    },
+    {
+        "id": 1187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001853.wav"
+            ]
+        }
+    },
+    {
+        "id": 1196,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000213.wav"
+            ]
+        }
+    },
+    {
+        "id": 1197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000422.wav"
+            ]
+        }
+    },
+    {
+        "id": 1203,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008377.wav"
+            ]
+        }
+    },
+    {
+        "id": 1206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004068.wav"
+            ]
+        }
+    },
+    {
+        "id": 1210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002258.wav"
+            ]
+        }
+    },
+    {
+        "id": 1219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004410.wav"
+            ]
+        }
+    },
+    {
+        "id": 1221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002060.wav"
+            ]
+        }
+    },
+    {
+        "id": 1223,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003847.wav"
+            ]
+        }
+    },
+    {
+        "id": 1227,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004132.wav"
+            ]
+        }
+    },
+    {
+        "id": 1232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001503.wav"
+            ]
+        }
+    },
+    {
+        "id": 1233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000694.wav"
+            ]
+        }
+    },
+    {
+        "id": 1237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003124.wav"
+            ]
+        }
+    },
+    {
+        "id": 1238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004202.wav"
+            ]
+        }
+    },
+    {
+        "id": 1250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007277.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007277.wav"
+            ]
+        }
+    },
+    {
+        "id": 1257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006789.wav"
+            ]
+        }
+    },
+    {
+        "id": 1269,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003262.wav"
+            ]
+        }
+    },
+    {
+        "id": 1272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007410.wav"
+            ]
+        }
+    },
+    {
+        "id": 1278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000516.wav"
+            ]
+        }
+    },
+    {
+        "id": 1280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005310.wav"
+            ]
+        }
+    },
+    {
+        "id": 1294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000208.wav"
+            ]
+        }
+    },
+    {
+        "id": 1305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000415.wav"
+            ]
+        }
+    },
+    {
+        "id": 1317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000522.wav"
+            ]
+        }
+    },
+    {
+        "id": 1323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007859.wav"
+            ]
+        }
+    },
+    {
+        "id": 1326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008414.wav"
+            ]
+        }
+    },
+    {
+        "id": 1327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005850.wav"
+            ]
+        }
+    },
+    {
+        "id": 1331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006412.wav"
+            ]
+        }
+    },
+    {
+        "id": 1343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002374.wav"
+            ]
+        }
+    },
+    {
+        "id": 1354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008133.wav"
+            ]
+        }
+    },
+    {
+        "id": 1365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006524.wav"
+            ]
+        }
+    },
+    {
+        "id": 1367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003868.wav"
+            ]
+        }
+    },
+    {
+        "id": 1370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004494.wav"
+            ]
+        }
+    },
+    {
+        "id": 1373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005691.wav"
+            ]
+        }
+    },
+    {
+        "id": 1375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008730.wav"
+            ]
+        }
+    },
+    {
+        "id": 1384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003079.wav"
+            ]
+        }
+    },
+    {
+        "id": 1388,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008543.wav"
+            ]
+        }
+    },
+    {
+        "id": 1390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003370.wav"
+            ]
+        }
+    },
+    {
+        "id": 1392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007364.wav"
+            ]
+        }
+    },
+    {
+        "id": 1393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004923.wav"
+            ]
+        }
+    },
+    {
+        "id": 1395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004283.wav"
+            ]
+        }
+    },
+    {
+        "id": 1403,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006549.wav"
+            ]
+        }
+    },
+    {
+        "id": 1413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001476.wav"
+            ]
+        }
+    },
+    {
+        "id": 1415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008051.wav"
+            ]
+        }
+    },
+    {
+        "id": 1417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004814.wav"
+            ]
+        }
+    },
+    {
+        "id": 1424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003731.wav"
+            ]
+        }
+    },
+    {
+        "id": 1427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001386.wav"
+            ]
+        }
+    },
+    {
+        "id": 1430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007893.wav"
+            ]
+        }
+    },
+    {
+        "id": 1431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006563.wav"
+            ]
+        }
+    },
+    {
+        "id": 1432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001533.wav"
+            ]
+        }
+    },
+    {
+        "id": 1440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005003.wav"
+            ]
+        }
+    },
+    {
+        "id": 1442,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002790.wav"
+            ]
+        }
+    },
+    {
+        "id": 1452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006211.wav"
+            ]
+        }
+    },
+    {
+        "id": 1453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006040.wav"
+            ]
+        }
+    },
+    {
+        "id": 1462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003643.wav"
+            ]
+        }
+    },
+    {
+        "id": 1464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008294.wav"
+            ]
+        }
+    },
+    {
+        "id": 1470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005535.wav"
+            ]
+        }
+    },
+    {
+        "id": 1475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004225.wav"
+            ]
+        }
+    },
+    {
+        "id": 1484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005755.wav"
+            ]
+        }
+    },
+    {
+        "id": 1487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001317.wav"
+            ]
+        }
+    },
+    {
+        "id": 1490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000309.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000309.wav"
+            ]
+        }
+    },
+    {
+        "id": 1514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001189.wav"
+            ]
+        }
+    },
+    {
+        "id": 1515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001182.wav"
+            ]
+        }
+    },
+    {
+        "id": 1529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002098.wav"
+            ]
+        }
+    },
+    {
+        "id": 1534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003392.wav"
+            ]
+        }
+    },
+    {
+        "id": 1538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001801.wav"
+            ]
+        }
+    },
+    {
+        "id": 1546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008087.wav"
+            ]
+        }
+    },
+    {
+        "id": 1551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002898.wav"
+            ]
+        }
+    },
+    {
+        "id": 1559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding electric_bass in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005807.wav"
+            ]
+        }
+    },
+    {
+        "id": 1562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001487.wav"
+            ]
+        }
+    },
+    {
+        "id": 1573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005500.wav"
+            ]
+        }
+    },
+    {
+        "id": 1574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000870.wav"
+            ]
+        }
+    },
+    {
+        "id": 1578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001717.wav"
+            ]
+        }
+    },
+    {
+        "id": 1580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000300.wav"
+            ]
+        }
+    },
+    {
+        "id": 1584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004828.wav"
+            ]
+        }
+    },
+    {
+        "id": 1596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000343.wav"
+            ]
+        }
+    },
+    {
+        "id": 1602,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005835.wav"
+            ]
+        }
+    },
+    {
+        "id": 1605,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007581.wav"
+            ]
+        }
+    },
+    {
+        "id": 1610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000059.wav"
+            ]
+        }
+    },
+    {
+        "id": 1615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007878.wav"
+            ]
+        }
+    },
+    {
+        "id": 1618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005262.wav"
+            ]
+        }
+    },
+    {
+        "id": 1625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006035.wav"
+            ]
+        }
+    },
+    {
+        "id": 1629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004229.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004229.wav"
+            ]
+        }
+    },
+    {
+        "id": 1636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006125.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006125.wav"
+            ]
+        }
+    },
+    {
+        "id": 1650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005196.wav"
+            ]
+        }
+    },
+    {
+        "id": 1658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000925.wav"
+            ]
+        }
+    },
+    {
+        "id": 1665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008820.wav"
+            ]
+        }
+    },
+    {
+        "id": 1685,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008786.wav"
+            ]
+        }
+    },
+    {
+        "id": 1702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003133.wav"
+            ]
+        }
+    },
+    {
+        "id": 1703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007001.wav"
+            ]
+        }
+    },
+    {
+        "id": 1705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003698.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003698.wav"
+            ]
+        }
+    },
+    {
+        "id": 1706,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005522.wav"
+            ]
+        }
+    },
+    {
+        "id": 1714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002666.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002666.wav"
+            ]
+        }
+    },
+    {
+        "id": 1727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001780.wav"
+            ]
+        }
+    },
+    {
+        "id": 1728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001341.wav"
+            ]
+        }
+    },
+    {
+        "id": 1733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007461.wav"
+            ]
+        }
+    },
+    {
+        "id": 1743,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001559.wav"
+            ]
+        }
+    },
+    {
+        "id": 1746,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000960.wav"
+            ]
+        }
+    },
+    {
+        "id": 1749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006532.wav"
+            ]
+        }
+    },
+    {
+        "id": 1752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002959.wav"
+            ]
+        }
+    },
+    {
+        "id": 1757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002610.wav"
+            ]
+        }
+    },
+    {
+        "id": 1758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006136.wav"
+            ]
+        }
+    },
+    {
+        "id": 1760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000015.wav"
+            ]
+        }
+    },
+    {
+        "id": 1775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001234.wav"
+            ]
+        }
+    },
+    {
+        "id": 1776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008458.wav"
+            ]
+        }
+    },
+    {
+        "id": 1813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004455.wav"
+            ]
+        }
+    },
+    {
+        "id": 1818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001183.wav"
+            ]
+        }
+    },
+    {
+        "id": 1822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000794.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000794.wav"
+            ]
+        }
+    },
+    {
+        "id": 1824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004193.wav"
+            ]
+        }
+    },
+    {
+        "id": 1826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002128.wav"
+            ]
+        }
+    },
+    {
+        "id": 1830,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001296.wav"
+            ]
+        }
+    },
+    {
+        "id": 1851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001080.wav"
+            ]
+        }
+    },
+    {
+        "id": 1858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002494.wav"
+            ]
+        }
+    },
+    {
+        "id": 1859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003196.wav"
+            ]
+        }
+    },
+    {
+        "id": 1868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006979.wav"
+            ]
+        }
+    },
+    {
+        "id": 1869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002553.wav"
+            ]
+        }
+    },
+    {
+        "id": 1875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004075.wav"
+            ]
+        }
+    },
+    {
+        "id": 1879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000723.wav"
+            ]
+        }
+    },
+    {
+        "id": 1886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007059.wav"
+            ]
+        }
+    },
+    {
+        "id": 1888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005680.wav"
+            ]
+        }
+    },
+    {
+        "id": 1892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000317.wav"
+            ]
+        }
+    },
+    {
+        "id": 1893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002468.wav"
+            ]
+        }
+    },
+    {
+        "id": 1900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000060.wav"
+            ]
+        }
+    },
+    {
+        "id": 1903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005876.wav"
+            ]
+        }
+    },
+    {
+        "id": 1918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003927.wav"
+            ]
+        }
+    },
+    {
+        "id": 1923,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005754.wav"
+            ]
+        }
+    },
+    {
+        "id": 1925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000738.wav"
+            ]
+        }
+    },
+    {
+        "id": 1928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007547.wav"
+            ]
+        }
+    },
+    {
+        "id": 1931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008825.wav"
+            ]
+        }
+    },
+    {
+        "id": 1951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000820.wav"
+            ]
+        }
+    },
+    {
+        "id": 1954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003996.wav"
+            ]
+        }
+    },
+    {
+        "id": 1955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006053.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006053.wav"
+            ]
+        }
+    },
+    {
+        "id": 1966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004010.wav"
+            ]
+        }
+    },
+    {
+        "id": 1969,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001682.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001682.wav"
+            ]
+        }
+    },
+    {
+        "id": 1977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008437.wav"
+            ]
+        }
+    },
+    {
+        "id": 1979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002394.wav"
+            ]
+        }
+    },
+    {
+        "id": 1982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008117.wav"
+            ]
+        }
+    },
+    {
+        "id": 1983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007792.wav"
+            ]
+        }
+    },
+    {
+        "id": 1991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006374.wav"
+            ]
+        }
+    },
+    {
+        "id": 1994,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005938.wav"
+            ]
+        }
+    },
+    {
+        "id": 1995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006485.wav"
+            ]
+        }
+    },
+    {
+        "id": 2000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001768.wav"
+            ]
+        }
+    },
+    {
+        "id": 2001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001573.wav"
+            ]
+        }
+    },
+    {
+        "id": 2015,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006916.wav"
+            ]
+        }
+    },
+    {
+        "id": 2021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008144.wav"
+            ]
+        }
+    },
+    {
+        "id": 2024,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005479.wav"
+            ]
+        }
+    },
+    {
+        "id": 2026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001542.wav"
+            ]
+        }
+    },
+    {
+        "id": 2032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008954.wav"
+            ]
+        }
+    },
+    {
+        "id": 2044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008352.wav"
+            ]
+        }
+    },
+    {
+        "id": 2047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000679.wav"
+            ]
+        }
+    },
+    {
+        "id": 2053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000787.wav"
+            ]
+        }
+    },
+    {
+        "id": 2057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008505.wav"
+            ]
+        }
+    },
+    {
+        "id": 2058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003509.wav"
+            ]
+        }
+    },
+    {
+        "id": 2067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001115.wav"
+            ]
+        }
+    },
+    {
+        "id": 2075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001788.wav"
+            ]
+        }
+    },
+    {
+        "id": 2082,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006358.wav"
+            ]
+        }
+    },
+    {
+        "id": 2085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005767.wav"
+            ]
+        }
+    },
+    {
+        "id": 2093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008677.wav"
+            ]
+        }
+    },
+    {
+        "id": 2101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002335.wav"
+            ]
+        }
+    },
+    {
+        "id": 2105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007104.wav"
+            ]
+        }
+    },
+    {
+        "id": 2107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002456.wav"
+            ]
+        }
+    },
+    {
+        "id": 2114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008805.wav"
+            ]
+        }
+    },
+    {
+        "id": 2128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007902.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007902.wav"
+            ]
+        }
+    },
+    {
+        "id": 2130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005743.wav"
+            ]
+        }
+    },
+    {
+        "id": 2133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001397.wav"
+            ]
+        }
+    },
+    {
+        "id": 2140,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008385.wav"
+            ]
+        }
+    },
+    {
+        "id": 2144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001668.wav"
+            ]
+        }
+    },
+    {
+        "id": 2151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005433.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005433.wav"
+            ]
+        }
+    },
+    {
+        "id": 2152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008812.wav"
+            ]
+        }
+    },
+    {
+        "id": 2159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001142.wav"
+            ]
+        }
+    },
+    {
+        "id": 2160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002244.wav"
+            ]
+        }
+    },
+    {
+        "id": 2162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007816.wav"
+            ]
+        }
+    },
+    {
+        "id": 2166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000279.wav"
+            ]
+        }
+    },
+    {
+        "id": 2167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002171.wav"
+            ]
+        }
+    },
+    {
+        "id": 2170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000875.wav"
+            ]
+        }
+    },
+    {
+        "id": 2173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007344.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007344.wav"
+            ]
+        }
+    },
+    {
+        "id": 2174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002741.wav"
+            ]
+        }
+    },
+    {
+        "id": 2178,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004035.wav"
+            ]
+        }
+    },
+    {
+        "id": 2180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding banjo in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007691.wav"
+            ]
+        }
+    },
+    {
+        "id": 2181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002559.wav"
+            ]
+        }
+    },
+    {
+        "id": 2188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002739.wav"
+            ]
+        }
+    },
+    {
+        "id": 2198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003170.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003170.wav"
+            ]
+        }
+    },
+    {
+        "id": 2199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006752.wav"
+            ]
+        }
+    },
+    {
+        "id": 2202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006778.wav"
+            ]
+        }
+    },
+    {
+        "id": 2204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002892.wav"
+            ]
+        }
+    },
+    {
+        "id": 2219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002164.wav"
+            ]
+        }
+    },
+    {
+        "id": 2220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007358.wav"
+            ]
+        }
+    },
+    {
+        "id": 2229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002805.wav"
+            ]
+        }
+    },
+    {
+        "id": 2240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000549.wav"
+            ]
+        }
+    },
+    {
+        "id": 2248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005089.wav"
+            ]
+        }
+    },
+    {
+        "id": 2249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007994.wav"
+            ]
+        }
+    },
+    {
+        "id": 2252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006276.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006276.wav"
+            ]
+        }
+    },
+    {
+        "id": 2254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007376.wav"
+            ]
+        }
+    },
+    {
+        "id": 2263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003974.wav"
+            ]
+        }
+    },
+    {
+        "id": 2266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007183.wav"
+            ]
+        }
+    },
+    {
+        "id": 2267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008733.wav"
+            ]
+        }
+    },
+    {
+        "id": 2280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004666.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004666.wav"
+            ]
+        }
+    },
+    {
+        "id": 2285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005128.wav"
+            ]
+        }
+    },
+    {
+        "id": 2292,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004653.wav"
+            ]
+        }
+    },
+    {
+        "id": 2293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006748.wav"
+            ]
+        }
+    },
+    {
+        "id": 2297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008011.wav"
+            ]
+        }
+    },
+    {
+        "id": 2299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005368.wav"
+            ]
+        }
+    },
+    {
+        "id": 2301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007679.wav"
+            ]
+        }
+    },
+    {
+        "id": 2312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008857.wav"
+            ]
+        }
+    },
+    {
+        "id": 2313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008073.wav"
+            ]
+        }
+    },
+    {
+        "id": 2321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007301.wav"
+            ]
+        }
+    },
+    {
+        "id": 2325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003175.wav"
+            ]
+        }
+    },
+    {
+        "id": 2327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006817.wav"
+            ]
+        }
+    },
+    {
+        "id": 2332,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003718.wav"
+            ]
+        }
+    },
+    {
+        "id": 2340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001717.wav"
+            ]
+        }
+    },
+    {
+        "id": 2344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006250.wav"
+            ]
+        }
+    },
+    {
+        "id": 2345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001894.wav"
+            ]
+        }
+    },
+    {
+        "id": 2346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001971.wav"
+            ]
+        }
+    },
+    {
+        "id": 2348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006957.wav"
+            ]
+        }
+    },
+    {
+        "id": 2350,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004354.wav"
+            ]
+        }
+    },
+    {
+        "id": 2353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008270.wav"
+            ]
+        }
+    },
+    {
+        "id": 2356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002071.wav"
+            ]
+        }
+    },
+    {
+        "id": 2361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003385.wav"
+            ]
+        }
+    },
+    {
+        "id": 2364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008161.wav"
+            ]
+        }
+    },
+    {
+        "id": 2368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008913.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008913.wav"
+            ]
+        }
+    },
+    {
+        "id": 2370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002823.wav"
+            ]
+        }
+    },
+    {
+        "id": 2377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007436.wav"
+            ]
+        }
+    },
+    {
+        "id": 2378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002731.wav"
+            ]
+        }
+    },
+    {
+        "id": 2380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000274.wav"
+            ]
+        }
+    },
+    {
+        "id": 2382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000944.wav"
+            ]
+        }
+    },
+    {
+        "id": 2384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000425.wav"
+            ]
+        }
+    },
+    {
+        "id": 2387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000816.wav"
+            ]
+        }
+    },
+    {
+        "id": 2389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006506.wav"
+            ]
+        }
+    },
+    {
+        "id": 2393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008016.wav"
+            ]
+        }
+    },
+    {
+        "id": 2397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008301.wav"
+            ]
+        }
+    },
+    {
+        "id": 2406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001505.wav"
+            ]
+        }
+    },
+    {
+        "id": 2428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001113.wav"
+            ]
+        }
+    },
+    {
+        "id": 2438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003490.wav"
+            ]
+        }
+    },
+    {
+        "id": 2440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002111.wav"
+            ]
+        }
+    },
+    {
+        "id": 2461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001901.wav"
+            ]
+        }
+    },
+    {
+        "id": 2462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001673.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001673.wav"
+            ]
+        }
+    },
+    {
+        "id": 2469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003100.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003100.wav"
+            ]
+        }
+    },
+    {
+        "id": 2470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006201.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006201.wav"
+            ]
+        }
+    },
+    {
+        "id": 2471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001649.wav"
+            ]
+        }
+    },
+    {
+        "id": 2480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000526.wav"
+            ]
+        }
+    },
+    {
+        "id": 2481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005249.wav"
+            ]
+        }
+    },
+    {
+        "id": 2483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004624.wav"
+            ]
+        }
+    },
+    {
+        "id": 2500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006500.wav"
+            ]
+        }
+    },
+    {
+        "id": 2501,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000003.wav"
+            ]
+        }
+    },
+    {
+        "id": 2507,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000385.wav"
+            ]
+        }
+    },
+    {
+        "id": 2516,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001910.wav"
+            ]
+        }
+    },
+    {
+        "id": 2518,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003563.wav"
+            ]
+        }
+    },
+    {
+        "id": 2520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000583.wav"
+            ]
+        }
+    },
+    {
+        "id": 2527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004389.wav"
+            ]
+        }
+    },
+    {
+        "id": 2531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002407.wav"
+            ]
+        }
+    },
+    {
+        "id": 2535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003085.wav"
+            ]
+        }
+    },
+    {
+        "id": 2537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008587.wav"
+            ]
+        }
+    },
+    {
+        "id": 2542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006803.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006803.wav"
+            ]
+        }
+    },
+    {
+        "id": 2555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005902.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005902.wav"
+            ]
+        }
+    },
+    {
+        "id": 2558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000133.wav"
+            ]
+        }
+    },
+    {
+        "id": 2569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008864.wav"
+            ]
+        }
+    },
+    {
+        "id": 2572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000506.wav"
+            ]
+        }
+    },
+    {
+        "id": 2575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001464.wav"
+            ]
+        }
+    },
+    {
+        "id": 2576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000688.wav"
+            ]
+        }
+    },
+    {
+        "id": 2587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002313.wav"
+            ]
+        }
+    },
+    {
+        "id": 2595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004060.wav"
+            ]
+        }
+    },
+    {
+        "id": 2603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003224.wav"
+            ]
+        }
+    },
+    {
+        "id": 2608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004524.wav"
+            ]
+        }
+    },
+    {
+        "id": 2609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001283.wav"
+            ]
+        }
+    },
+    {
+        "id": 2610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001220.wav"
+            ]
+        }
+    },
+    {
+        "id": 2614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000408.wav"
+            ]
+        }
+    },
+    {
+        "id": 2616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002044.wav"
+            ]
+        }
+    },
+    {
+        "id": 2620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001442.wav"
+            ]
+        }
+    },
+    {
+        "id": 2621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008154.wav"
+            ]
+        }
+    },
+    {
+        "id": 2628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004598.wav"
+            ]
+        }
+    },
+    {
+        "id": 2633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006923.wav"
+            ]
+        }
+    },
+    {
+        "id": 2646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000648.wav"
+            ]
+        }
+    },
+    {
+        "id": 2649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004120.wav"
+            ]
+        }
+    },
+    {
+        "id": 2653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003922.wav"
+            ]
+        }
+    },
+    {
+        "id": 2656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004988.wav"
+            ]
+        }
+    },
+    {
+        "id": 2658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007269.wav"
+            ]
+        }
+    },
+    {
+        "id": 2667,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004840.wav"
+            ]
+        }
+    },
+    {
+        "id": 2670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008916.wav"
+            ]
+        }
+    },
+    {
+        "id": 2675,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007630.wav"
+            ]
+        }
+    },
+    {
+        "id": 2678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008076.wav"
+            ]
+        }
+    },
+    {
+        "id": 2679,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003436.wav"
+            ]
+        }
+    },
+    {
+        "id": 2681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003072.wav"
+            ]
+        }
+    },
+    {
+        "id": 2682,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000704.wav"
+            ]
+        }
+    },
+    {
+        "id": 2684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001902.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001902.wav"
+            ]
+        }
+    },
+    {
+        "id": 2686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004373.wav"
+            ]
+        }
+    },
+    {
+        "id": 2687,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004578.wav"
+            ]
+        }
+    },
+    {
+        "id": 2693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006878.wav"
+            ]
+        }
+    },
+    {
+        "id": 2695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003452.wav"
+            ]
+        }
+    },
+    {
+        "id": 2707,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005269.wav"
+            ]
+        }
+    },
+    {
+        "id": 2722,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004918.wav"
+            ]
+        }
+    },
+    {
+        "id": 2724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007763.wav"
+            ]
+        }
+    },
+    {
+        "id": 2730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001338.wav"
+            ]
+        }
+    },
+    {
+        "id": 2748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001290.wav"
+            ]
+        }
+    },
+    {
+        "id": 2757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001012.wav"
+            ]
+        }
+    },
+    {
+        "id": 2766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002671.wav"
+            ]
+        }
+    },
+    {
+        "id": 2768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008759.wav"
+            ]
+        }
+    },
+    {
+        "id": 2771,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001046.wav"
+            ]
+        }
+    },
+    {
+        "id": 2784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000830.wav"
+            ]
+        }
+    },
+    {
+        "id": 2795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002946.wav"
+            ]
+        }
+    },
+    {
+        "id": 2797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002649.wav"
+            ]
+        }
+    },
+    {
+        "id": 2800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005735.wav"
+            ]
+        }
+    },
+    {
+        "id": 2802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008627.wav"
+            ]
+        }
+    },
+    {
+        "id": 2804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002665.wav"
+            ]
+        }
+    },
+    {
+        "id": 2812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008021.wav"
+            ]
+        }
+    },
+    {
+        "id": 2816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002249.wav"
+            ]
+        }
+    },
+    {
+        "id": 2818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002143.wav"
+            ]
+        }
+    },
+    {
+        "id": 2821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003958.wav"
+            ]
+        }
+    },
+    {
+        "id": 2831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006955.wav"
+            ]
+        }
+    },
+    {
+        "id": 2835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008688.wav"
+            ]
+        }
+    },
+    {
+        "id": 2851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006171.wav"
+            ]
+        }
+    },
+    {
+        "id": 2866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003067.wav"
+            ]
+        }
+    },
+    {
+        "id": 2872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000253.wav"
+            ]
+        }
+    },
+    {
+        "id": 2874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007185.wav"
+            ]
+        }
+    },
+    {
+        "id": 2892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003116.wav"
+            ]
+        }
+    },
+    {
+        "id": 2894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005280.wav"
+            ]
+        }
+    },
+    {
+        "id": 2900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004642.wav"
+            ]
+        }
+    },
+    {
+        "id": 2909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003452.wav"
+            ]
+        }
+    },
+    {
+        "id": 2911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008756.wav"
+            ]
+        }
+    },
+    {
+        "id": 2912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008918.wav"
+            ]
+        }
+    },
+    {
+        "id": 2914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003158.wav"
+            ]
+        }
+    },
+    {
+        "id": 2919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004099.wav"
+            ]
+        }
+    },
+    {
+        "id": 2922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001550.wav"
+            ]
+        }
+    },
+    {
+        "id": 2927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003648.wav"
+            ]
+        }
+    },
+    {
+        "id": 2931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001158.wav"
+            ]
+        }
+    },
+    {
+        "id": 2932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006321.wav"
+            ]
+        }
+    },
+    {
+        "id": 2937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004552.wav"
+            ]
+        }
+    },
+    {
+        "id": 2939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002335.wav"
+            ]
+        }
+    },
+    {
+        "id": 2944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000441.wav"
+            ]
+        }
+    },
+    {
+        "id": 2945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000823.wav"
+            ]
+        }
+    },
+    {
+        "id": 2946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003470.wav"
+            ]
+        }
+    },
+    {
+        "id": 2953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004562.wav"
+            ]
+        }
+    },
+    {
+        "id": 2954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003304.wav"
+            ]
+        }
+    },
+    {
+        "id": 2958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005973.wav"
+            ]
+        }
+    },
+    {
+        "id": 2961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003997.wav"
+            ]
+        }
+    },
+    {
+        "id": 2966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003467.wav"
+            ]
+        }
+    },
+    {
+        "id": 2973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001985.wav"
+            ]
+        }
+    },
+    {
+        "id": 2982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008178.wav"
+            ]
+        }
+    },
+    {
+        "id": 2985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000944.wav"
+            ]
+        }
+    },
+    {
+        "id": 2990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008689.wav"
+            ]
+        }
+    },
+    {
+        "id": 3004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003187.wav"
+            ]
+        }
+    },
+    {
+        "id": 3009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000012.wav"
+            ]
+        }
+    },
+    {
+        "id": 3010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008967.wav"
+            ]
+        }
+    },
+    {
+        "id": 3012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001335.wav"
+            ]
+        }
+    },
+    {
+        "id": 3016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002879.wav"
+            ]
+        }
+    },
+    {
+        "id": 3019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008790.wav"
+            ]
+        }
+    },
+    {
+        "id": 3022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006675.wav"
+            ]
+        }
+    },
+    {
+        "id": 3027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007224.wav"
+            ]
+        }
+    },
+    {
+        "id": 3028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001666.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001666.wav"
+            ]
+        }
+    },
+    {
+        "id": 3032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004813.wav"
+            ]
+        }
+    },
+    {
+        "id": 3034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008241.wav"
+            ]
+        }
+    },
+    {
+        "id": 3035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005023.wav"
+            ]
+        }
+    },
+    {
+        "id": 3050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006986.wav"
+            ]
+        }
+    },
+    {
+        "id": 3056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006325.wav"
+            ]
+        }
+    },
+    {
+        "id": 3074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004222.wav"
+            ]
+        }
+    },
+    {
+        "id": 3085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001940.wav"
+            ]
+        }
+    },
+    {
+        "id": 3086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007124.wav"
+            ]
+        }
+    },
+    {
+        "id": 3089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005550.wav"
+            ]
+        }
+    },
+    {
+        "id": 3093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007113.wav"
+            ]
+        }
+    },
+    {
+        "id": 3094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007952.wav"
+            ]
+        }
+    },
+    {
+        "id": 3097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001586.wav"
+            ]
+        }
+    },
+    {
+        "id": 3098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005370.wav"
+            ]
+        }
+    },
+    {
+        "id": 3111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006948.wav"
+            ]
+        }
+    },
+    {
+        "id": 3114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005430.wav"
+            ]
+        }
+    },
+    {
+        "id": 3120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005298.wav"
+            ]
+        }
+    },
+    {
+        "id": 3125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004327.wav"
+            ]
+        }
+    },
+    {
+        "id": 3143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004264.wav"
+            ]
+        }
+    },
+    {
+        "id": 3145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007269.wav"
+            ]
+        }
+    },
+    {
+        "id": 3150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003949.wav"
+            ]
+        }
+    },
+    {
+        "id": 3155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007513.wav"
+            ]
+        }
+    },
+    {
+        "id": 3156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001596.wav"
+            ]
+        }
+    },
+    {
+        "id": 3159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007349.wav"
+            ]
+        }
+    },
+    {
+        "id": 3166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006840.wav"
+            ]
+        }
+    },
+    {
+        "id": 3171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006638.wav"
+            ]
+        }
+    },
+    {
+        "id": 3174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006906.wav"
+            ]
+        }
+    },
+    {
+        "id": 3191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008689.wav"
+            ]
+        }
+    },
+    {
+        "id": 3205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005840.wav"
+            ]
+        }
+    },
+    {
+        "id": 3212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007831.wav"
+            ]
+        }
+    },
+    {
+        "id": 3213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001073.wav"
+            ]
+        }
+    },
+    {
+        "id": 3217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006190.wav"
+            ]
+        }
+    },
+    {
+        "id": 3221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003249.wav"
+            ]
+        }
+    },
+    {
+        "id": 3229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005853.wav"
+            ]
+        }
+    },
+    {
+        "id": 3230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007665.wav"
+            ]
+        }
+    },
+    {
+        "id": 3243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003358.wav"
+            ]
+        }
+    },
+    {
+        "id": 3244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006279.wav"
+            ]
+        }
+    },
+    {
+        "id": 3247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001955.wav"
+            ]
+        }
+    },
+    {
+        "id": 3252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000269.wav"
+            ]
+        }
+    },
+    {
+        "id": 3254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004135.wav"
+            ]
+        }
+    },
+    {
+        "id": 3260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005710.wav"
+            ]
+        }
+    },
+    {
+        "id": 3266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007335.wav"
+            ]
+        }
+    },
+    {
+        "id": 3268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008644.wav"
+            ]
+        }
+    },
+    {
+        "id": 3270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008228.wav"
+            ]
+        }
+    },
+    {
+        "id": 3273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006157.wav"
+            ]
+        }
+    },
+    {
+        "id": 3279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000670.wav"
+            ]
+        }
+    },
+    {
+        "id": 3280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001753.wav"
+            ]
+        }
+    },
+    {
+        "id": 3283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004345.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004345.wav"
+            ]
+        }
+    },
+    {
+        "id": 3285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004877.wav"
+            ]
+        }
+    },
+    {
+        "id": 3287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003537.wav"
+            ]
+        }
+    },
+    {
+        "id": 3293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005922.wav"
+            ]
+        }
+    },
+    {
+        "id": 3299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001102.wav"
+            ]
+        }
+    },
+    {
+        "id": 3304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006017.wav"
+            ]
+        }
+    },
+    {
+        "id": 3322,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006751.wav"
+            ]
+        }
+    },
+    {
+        "id": 3324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004429.wav"
+            ]
+        }
+    },
+    {
+        "id": 3327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003776.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003776.wav"
+            ]
+        }
+    },
+    {
+        "id": 3335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003576.wav"
+            ]
+        }
+    },
+    {
+        "id": 3345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008210.wav"
+            ]
+        }
+    },
+    {
+        "id": 3358,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008109.wav"
+            ]
+        }
+    },
+    {
+        "id": 3361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006368.wav"
+            ]
+        }
+    },
+    {
+        "id": 3368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001015.wav"
+            ]
+        }
+    },
+    {
+        "id": 3388,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004176.wav"
+            ]
+        }
+    },
+    {
+        "id": 3395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000706.wav"
+            ]
+        }
+    },
+    {
+        "id": 3397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007307.wav"
+            ]
+        }
+    },
+    {
+        "id": 3399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005548.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005548.wav"
+            ]
+        }
+    },
+    {
+        "id": 3402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002636.wav"
+            ]
+        }
+    },
+    {
+        "id": 3404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000924.wav"
+            ]
+        }
+    },
+    {
+        "id": 3406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003999.wav"
+            ]
+        }
+    },
+    {
+        "id": 3411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007523.wav"
+            ]
+        }
+    },
+    {
+        "id": 3412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001914.wav"
+            ]
+        }
+    },
+    {
+        "id": 3415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003623.wav"
+            ]
+        }
+    },
+    {
+        "id": 3416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003014.wav"
+            ]
+        }
+    },
+    {
+        "id": 3428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008199.wav"
+            ]
+        }
+    },
+    {
+        "id": 3438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004722.wav"
+            ]
+        }
+    },
+    {
+        "id": 3447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006793.wav"
+            ]
+        }
+    },
+    {
+        "id": 3457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005058.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005058.wav"
+            ]
+        }
+    },
+    {
+        "id": 3463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000501.wav"
+            ]
+        }
+    },
+    {
+        "id": 3472,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005030.wav"
+            ]
+        }
+    },
+    {
+        "id": 3476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005790.wav"
+            ]
+        }
+    },
+    {
+        "id": 3489,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002342.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002342.wav"
+            ]
+        }
+    },
+    {
+        "id": 3497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004247.wav"
+            ]
+        }
+    },
+    {
+        "id": 3503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001142.wav"
+            ]
+        }
+    },
+    {
+        "id": 3508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006239.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006239.wav"
+            ]
+        }
+    },
+    {
+        "id": 3510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000719.wav"
+            ]
+        }
+    },
+    {
+        "id": 3515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003940.wav"
+            ]
+        }
+    },
+    {
+        "id": 3520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001317.wav"
+            ]
+        }
+    },
+    {
+        "id": 3527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001690.wav"
+            ]
+        }
+    },
+    {
+        "id": 3531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002578.wav"
+            ]
+        }
+    },
+    {
+        "id": 3533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005868.wav"
+            ]
+        }
+    },
+    {
+        "id": 3534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003023.wav"
+            ]
+        }
+    },
+    {
+        "id": 3536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005182.wav"
+            ]
+        }
+    },
+    {
+        "id": 3540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004820.wav"
+            ]
+        }
+    },
+    {
+        "id": 3541,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001249.wav"
+            ]
+        }
+    },
+    {
+        "id": 3542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002658.wav"
+            ]
+        }
+    },
+    {
+        "id": 3545,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006261.wav"
+            ]
+        }
+    },
+    {
+        "id": 3561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008128.wav"
+            ]
+        }
+    },
+    {
+        "id": 3564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007159.wav"
+            ]
+        }
+    },
+    {
+        "id": 3567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001311.wav"
+            ]
+        }
+    },
+    {
+        "id": 3568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005290.wav"
+            ]
+        }
+    },
+    {
+        "id": 3575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000494.wav"
+            ]
+        }
+    },
+    {
+        "id": 3578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001617.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001617.wav"
+            ]
+        }
+    },
+    {
+        "id": 3582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002336.wav"
+            ]
+        }
+    },
+    {
+        "id": 3583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005450.wav"
+            ]
+        }
+    },
+    {
+        "id": 3584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006935.wav"
+            ]
+        }
+    },
+    {
+        "id": 3585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008832.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008832.wav"
+            ]
+        }
+    },
+    {
+        "id": 3596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000290.wav"
+            ]
+        }
+    },
+    {
+        "id": 3597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003107.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003107.wav"
+            ]
+        }
+    },
+    {
+        "id": 3599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006292.wav"
+            ]
+        }
+    },
+    {
+        "id": 3607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007004.wav"
+            ]
+        }
+    },
+    {
+        "id": 3608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008659.wav"
+            ]
+        }
+    },
+    {
+        "id": 3610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005082.wav"
+            ]
+        }
+    },
+    {
+        "id": 3611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004291.wav"
+            ]
+        }
+    },
+    {
+        "id": 3628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002115.wav"
+            ]
+        }
+    },
+    {
+        "id": 3633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004751.wav"
+            ]
+        }
+    },
+    {
+        "id": 3634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008328.wav"
+            ]
+        }
+    },
+    {
+        "id": 3635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002869.wav"
+            ]
+        }
+    },
+    {
+        "id": 3636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001158.wav"
+            ]
+        }
+    },
+    {
+        "id": 3650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007012.wav"
+            ]
+        }
+    },
+    {
+        "id": 3668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005126.wav"
+            ]
+        }
+    },
+    {
+        "id": 3686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004403.wav"
+            ]
+        }
+    },
+    {
+        "id": 3691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001799.wav"
+            ]
+        }
+    },
+    {
+        "id": 3697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002912.wav"
+            ]
+        }
+    },
+    {
+        "id": 3701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002733.wav"
+            ]
+        }
+    },
+    {
+        "id": 3703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003868.wav"
+            ]
+        }
+    },
+    {
+        "id": 3714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006861.wav"
+            ]
+        }
+    },
+    {
+        "id": 3721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002504.wav"
+            ]
+        }
+    },
+    {
+        "id": 3737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002318.wav"
+            ]
+        }
+    },
+    {
+        "id": 3742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005172.wav"
+            ]
+        }
+    },
+    {
+        "id": 3753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000907.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000907.wav"
+            ]
+        }
+    },
+    {
+        "id": 3759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004864.wav"
+            ]
+        }
+    },
+    {
+        "id": 3763,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008241.wav"
+            ]
+        }
+    },
+    {
+        "id": 3774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002208.wav"
+            ]
+        }
+    },
+    {
+        "id": 3777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001700.wav"
+            ]
+        }
+    },
+    {
+        "id": 3786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004739.wav"
+            ]
+        }
+    },
+    {
+        "id": 3787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002021.wav"
+            ]
+        }
+    },
+    {
+        "id": 3790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007409.wav"
+            ]
+        }
+    },
+    {
+        "id": 3791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000475.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000475.wav"
+            ]
+        }
+    },
+    {
+        "id": 3797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004976.wav"
+            ]
+        }
+    },
+    {
+        "id": 3798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007173.wav"
+            ]
+        }
+    },
+    {
+        "id": 3799,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000390.wav"
+            ]
+        }
+    },
+    {
+        "id": 3802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000054.wav"
+            ]
+        }
+    },
+    {
+        "id": 3803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002904.wav"
+            ]
+        }
+    },
+    {
+        "id": 3809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005555.wav"
+            ]
+        }
+    },
+    {
+        "id": 3810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007965.wav"
+            ]
+        }
+    },
+    {
+        "id": 3811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002205.wav"
+            ]
+        }
+    },
+    {
+        "id": 3820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001592.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001592.wav"
+            ]
+        }
+    },
+    {
+        "id": 3823,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001208.wav"
+            ]
+        }
+    },
+    {
+        "id": 3835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007050.wav"
+            ]
+        }
+    },
+    {
+        "id": 3837,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003552.wav"
+            ]
+        }
+    },
+    {
+        "id": 3846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005398.wav"
+            ]
+        }
+    },
+    {
+        "id": 3853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004161.wav"
+            ]
+        }
+    },
+    {
+        "id": 3854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005440.wav"
+            ]
+        }
+    },
+    {
+        "id": 3857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007652.wav"
+            ]
+        }
+    },
+    {
+        "id": 3863,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005202.wav"
+            ]
+        }
+    },
+    {
+        "id": 3865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003949.wav"
+            ]
+        }
+    },
+    {
+        "id": 3870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004429.wav"
+            ]
+        }
+    },
+    {
+        "id": 3871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007806.wav"
+            ]
+        }
+    },
+    {
+        "id": 3872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007998.wav"
+            ]
+        }
+    },
+    {
+        "id": 3874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003288.wav"
+            ]
+        }
+    },
+    {
+        "id": 3877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001424.wav"
+            ]
+        }
+    },
+    {
+        "id": 3882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008791.wav"
+            ]
+        }
+    },
+    {
+        "id": 3884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002004.wav"
+            ]
+        }
+    },
+    {
+        "id": 3886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002240.wav"
+            ]
+        }
+    },
+    {
+        "id": 3889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007152.wav"
+            ]
+        }
+    },
+    {
+        "id": 3896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003458.wav"
+            ]
+        }
+    },
+    {
+        "id": 3900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004604.wav"
+            ]
+        }
+    },
+    {
+        "id": 3917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005230.wav"
+            ]
+        }
+    },
+    {
+        "id": 3920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002190.wav"
+            ]
+        }
+    },
+    {
+        "id": 3934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004506.wav"
+            ]
+        }
+    },
+    {
+        "id": 3936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003708.wav"
+            ]
+        }
+    },
+    {
+        "id": 3939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003653.wav"
+            ]
+        }
+    },
+    {
+        "id": 3941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004464.wav"
+            ]
+        }
+    },
+    {
+        "id": 3952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002951.wav"
+            ]
+        }
+    },
+    {
+        "id": 3953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001749.wav"
+            ]
+        }
+    },
+    {
+        "id": 3954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003859.wav"
+            ]
+        }
+    },
+    {
+        "id": 3956,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003912.wav"
+            ]
+        }
+    },
+    {
+        "id": 3958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007652.wav"
+            ]
+        }
+    },
+    {
+        "id": 3959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006933.wav"
+            ]
+        }
+    },
+    {
+        "id": 3961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001705.wav"
+            ]
+        }
+    },
+    {
+        "id": 3962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005570.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005570.wav"
+            ]
+        }
+    },
+    {
+        "id": 3965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002835.wav"
+            ]
+        }
+    },
+    {
+        "id": 3969,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005948.wav"
+            ]
+        }
+    },
+    {
+        "id": 3975,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005093.wav"
+            ]
+        }
+    },
+    {
+        "id": 3978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000067.wav"
+            ]
+        }
+    },
+    {
+        "id": 3983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003818.wav"
+            ]
+        }
+    },
+    {
+        "id": 3986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006297.wav"
+            ]
+        }
+    },
+    {
+        "id": 3989,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005836.wav"
+            ]
+        }
+    },
+    {
+        "id": 4005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002439.wav"
+            ]
+        }
+    },
+    {
+        "id": 4009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007202.wav"
+            ]
+        }
+    },
+    {
+        "id": 4018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000732.wav"
+            ]
+        }
+    },
+    {
+        "id": 4030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003859.wav"
+            ]
+        }
+    },
+    {
+        "id": 4038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005757.wav"
+            ]
+        }
+    },
+    {
+        "id": 4047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004606.wav"
+            ]
+        }
+    },
+    {
+        "id": 4054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000247.wav"
+            ]
+        }
+    },
+    {
+        "id": 4061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002476.wav"
+            ]
+        }
+    },
+    {
+        "id": 4073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001068.wav"
+            ]
+        }
+    },
+    {
+        "id": 4075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006153.wav"
+            ]
+        }
+    },
+    {
+        "id": 4079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002653.wav"
+            ]
+        }
+    },
+    {
+        "id": 4080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003368.wav"
+            ]
+        }
+    },
+    {
+        "id": 4081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001250.wav"
+            ]
+        }
+    },
+    {
+        "id": 4087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000655.wav"
+            ]
+        }
+    },
+    {
+        "id": 4093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003795.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003795.wav"
+            ]
+        }
+    },
+    {
+        "id": 4095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007220.wav"
+            ]
+        }
+    },
+    {
+        "id": 4097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003759.wav"
+            ]
+        }
+    },
+    {
+        "id": 4099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007990.wav"
+            ]
+        }
+    },
+    {
+        "id": 4112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004974.wav"
+            ]
+        }
+    },
+    {
+        "id": 4116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004075.wav"
+            ]
+        }
+    },
+    {
+        "id": 4121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments sound in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003402.wav"
+            ]
+        }
+    },
+    {
+        "id": 4127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003788.wav"
+            ]
+        }
+    },
+    {
+        "id": 4128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003367.wav"
+            ]
+        }
+    },
+    {
+        "id": 4143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left more rhythmic than the banjo on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006438.wav"
+            ]
+        }
+    },
+    {
+        "id": 4146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005563.wav"
+            ]
+        }
+    },
+    {
+        "id": 4164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005077.wav"
+            ]
+        }
+    },
+    {
+        "id": 4165,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001608.wav"
+            ]
+        }
+    },
+    {
+        "id": 4169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002323.wav"
+            ]
+        }
+    },
+    {
+        "id": 4173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007608.wav"
+            ]
+        }
+    },
+    {
+        "id": 4178,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005018.wav"
+            ]
+        }
+    },
+    {
+        "id": 4182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007944.wav"
+            ]
+        }
+    },
+    {
+        "id": 4183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005251.wav"
+            ]
+        }
+    },
+    {
+        "id": 4190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003253.wav"
+            ]
+        }
+    },
+    {
+        "id": 4207,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001430.wav"
+            ]
+        }
+    },
+    {
+        "id": 4212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007839.wav"
+            ]
+        }
+    },
+    {
+        "id": 4217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005650.wav"
+            ]
+        }
+    },
+    {
+        "id": 4222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007911.wav"
+            ]
+        }
+    },
+    {
+        "id": 4230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008787.wav"
+            ]
+        }
+    },
+    {
+        "id": 4233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006817.wav"
+            ]
+        }
+    },
+    {
+        "id": 4236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008550.wav"
+            ]
+        }
+    },
+    {
+        "id": 4240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003119.wav"
+            ]
+        }
+    },
+    {
+        "id": 4243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008726.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008726.wav"
+            ]
+        }
+    },
+    {
+        "id": 4250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005640.wav"
+            ]
+        }
+    },
+    {
+        "id": 4252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000159.wav"
+            ]
+        }
+    },
+    {
+        "id": 4264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000727.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000727.wav"
+            ]
+        }
+    },
+    {
+        "id": 4272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008090.wav"
+            ]
+        }
+    },
+    {
+        "id": 4277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007426.wav"
+            ]
+        }
+    },
+    {
+        "id": 4281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001427.wav"
+            ]
+        }
+    },
+    {
+        "id": 4286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002759.wav"
+            ]
+        }
+    },
+    {
+        "id": 4291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007465.wav"
+            ]
+        }
+    },
+    {
+        "id": 4294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000825.wav"
+            ]
+        }
+    },
+    {
+        "id": 4304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007094.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007094.wav"
+            ]
+        }
+    },
+    {
+        "id": 4305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004842.wav"
+            ]
+        }
+    },
+    {
+        "id": 4314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008954.wav"
+            ]
+        }
+    },
+    {
+        "id": 4317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007634.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007634.wav"
+            ]
+        }
+    },
+    {
+        "id": 4320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006642.wav"
+            ]
+        }
+    },
+    {
+        "id": 4323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006861.wav"
+            ]
+        }
+    },
+    {
+        "id": 4324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005989.wav"
+            ]
+        }
+    },
+    {
+        "id": 4325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006710.wav"
+            ]
+        }
+    },
+    {
+        "id": 4338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006950.wav"
+            ]
+        }
+    },
+    {
+        "id": 4341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008839.wav"
+            ]
+        }
+    },
+    {
+        "id": 4342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008126.wav"
+            ]
+        }
+    },
+    {
+        "id": 4359,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006001.wav"
+            ]
+        }
+    },
+    {
+        "id": 4361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001539.wav"
+            ]
+        }
+    },
+    {
+        "id": 4370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004476.wav"
+            ]
+        }
+    },
+    {
+        "id": 4376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003751.wav"
+            ]
+        }
+    },
+    {
+        "id": 4387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004266.wav"
+            ]
+        }
+    },
+    {
+        "id": 4400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008161.wav"
+            ]
+        }
+    },
+    {
+        "id": 4419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004676.wav"
+            ]
+        }
+    },
+    {
+        "id": 4429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008910.wav"
+            ]
+        }
+    },
+    {
+        "id": 4431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000999.wav"
+            ]
+        }
+    },
+    {
+        "id": 4436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007880.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007880.wav"
+            ]
+        }
+    },
+    {
+        "id": 4441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005616.wav"
+            ]
+        }
+    },
+    {
+        "id": 4444,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003427.wav"
+            ]
+        }
+    },
+    {
+        "id": 4445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007559.wav"
+            ]
+        }
+    },
+    {
+        "id": 4452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003161.wav"
+            ]
+        }
+    },
+    {
+        "id": 4457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006875.wav"
+            ]
+        }
+    },
+    {
+        "id": 4461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000451.wav"
+            ]
+        }
+    },
+    {
+        "id": 4464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000868.wav"
+            ]
+        }
+    },
+    {
+        "id": 4469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003103.wav"
+            ]
+        }
+    },
+    {
+        "id": 4472,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006316.wav"
+            ]
+        }
+    },
+    {
+        "id": 4476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006771.wav"
+            ]
+        }
+    },
+    {
+        "id": 4481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005847.wav"
+            ]
+        }
+    },
+    {
+        "id": 4482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005979.wav"
+            ]
+        }
+    },
+    {
+        "id": 4491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006238.wav"
+            ]
+        }
+    },
+    {
+        "id": 4501,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006459.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006459.wav"
+            ]
+        }
+    },
+    {
+        "id": 4502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005493.wav"
+            ]
+        }
+    },
+    {
+        "id": 4504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006392.wav"
+            ]
+        }
+    },
+    {
+        "id": 4506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005704.wav"
+            ]
+        }
+    },
+    {
+        "id": 4510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000774.wav"
+            ]
+        }
+    },
+    {
+        "id": 4520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006105.wav"
+            ]
+        }
+    },
+    {
+        "id": 4529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008615.wav"
+            ]
+        }
+    },
+    {
+        "id": 4539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002847.wav"
+            ]
+        }
+    },
+    {
+        "id": 4540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005795.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005795.wav"
+            ]
+        }
+    },
+    {
+        "id": 4545,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007070.wav"
+            ]
+        }
+    },
+    {
+        "id": 4546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008130.wav"
+            ]
+        }
+    },
+    {
+        "id": 4547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001331.wav"
+            ]
+        }
+    },
+    {
+        "id": 4558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001813.wav"
+            ]
+        }
+    },
+    {
+        "id": 4573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002079.wav"
+            ]
+        }
+    },
+    {
+        "id": 4574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000917.wav"
+            ]
+        }
+    },
+    {
+        "id": 4579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000823.wav"
+            ]
+        }
+    },
+    {
+        "id": 4584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001858.wav"
+            ]
+        }
+    },
+    {
+        "id": 4591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007706.wav"
+            ]
+        }
+    },
+    {
+        "id": 4599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000844.wav"
+            ]
+        }
+    },
+    {
+        "id": 4608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004735.wav"
+            ]
+        }
+    },
+    {
+        "id": 4610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001766.wav"
+            ]
+        }
+    },
+    {
+        "id": 4616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005013.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005013.wav"
+            ]
+        }
+    },
+    {
+        "id": 4622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006814.wav"
+            ]
+        }
+    },
+    {
+        "id": 4626,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004001.wav"
+            ]
+        }
+    },
+    {
+        "id": 4627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001704.wav"
+            ]
+        }
+    },
+    {
+        "id": 4635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001830.wav"
+            ]
+        }
+    },
+    {
+        "id": 4637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003056.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003056.wav"
+            ]
+        }
+    },
+    {
+        "id": 4639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006835.wav"
+            ]
+        }
+    },
+    {
+        "id": 4641,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007717.wav"
+            ]
+        }
+    },
+    {
+        "id": 4643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004853.wav"
+            ]
+        }
+    },
+    {
+        "id": 4651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000108.wav"
+            ]
+        }
+    },
+    {
+        "id": 4654,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007835.wav"
+            ]
+        }
+    },
+    {
+        "id": 4665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003232.wav"
+            ]
+        }
+    },
+    {
+        "id": 4670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000028.wav"
+            ]
+        }
+    },
+    {
+        "id": 4676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000114.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000114.wav"
+            ]
+        }
+    },
+    {
+        "id": 4677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000053.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000053.wav"
+            ]
+        }
+    },
+    {
+        "id": 4683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006190.wav"
+            ]
+        }
+    },
+    {
+        "id": 4689,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008969.wav"
+            ]
+        }
+    },
+    {
+        "id": 4690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005396.wav"
+            ]
+        }
+    },
+    {
+        "id": 4692,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000778.wav"
+            ]
+        }
+    },
+    {
+        "id": 4701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001841.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001841.wav"
+            ]
+        }
+    },
+    {
+        "id": 4714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005242.wav"
+            ]
+        }
+    },
+    {
+        "id": 4719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008121.wav"
+            ]
+        }
+    },
+    {
+        "id": 4721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004508.wav"
+            ]
+        }
+    },
+    {
+        "id": 4727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007468.wav"
+            ]
+        }
+    },
+    {
+        "id": 4728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002128.wav"
+            ]
+        }
+    },
+    {
+        "id": 4730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000489.wav"
+            ]
+        }
+    },
+    {
+        "id": 4732,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001361.wav"
+            ]
+        }
+    },
+    {
+        "id": 4734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005955.wav"
+            ]
+        }
+    },
+    {
+        "id": 4735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003526.wav"
+            ]
+        }
+    },
+    {
+        "id": 4736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005637.wav"
+            ]
+        }
+    },
+    {
+        "id": 4740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001450.wav"
+            ]
+        }
+    },
+    {
+        "id": 4753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002117.wav"
+            ]
+        }
+    },
+    {
+        "id": 4756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003722.wav"
+            ]
+        }
+    },
+    {
+        "id": 4767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003132.wav"
+            ]
+        }
+    },
+    {
+        "id": 4770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007364.wav"
+            ]
+        }
+    },
+    {
+        "id": 4775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000234.wav"
+            ]
+        }
+    },
+    {
+        "id": 4797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006708.wav"
+            ]
+        }
+    },
+    {
+        "id": 4798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006589.wav"
+            ]
+        }
+    },
+    {
+        "id": 4801,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006071.wav"
+            ]
+        }
+    },
+    {
+        "id": 4812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003037.wav"
+            ]
+        }
+    },
+    {
+        "id": 4815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007143.wav"
+            ]
+        }
+    },
+    {
+        "id": 4828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008324.wav"
+            ]
+        }
+    },
+    {
+        "id": 4829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004774.wav"
+            ]
+        }
+    },
+    {
+        "id": 4830,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000901.wav"
+            ]
+        }
+    },
+    {
+        "id": 4831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002836.wav"
+            ]
+        }
+    },
+    {
+        "id": 4832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the right more rhythmic than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001073.wav"
+            ]
+        }
+    },
+    {
+        "id": 4838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008016.wav"
+            ]
+        }
+    },
+    {
+        "id": 4839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003454.wav"
+            ]
+        }
+    },
+    {
+        "id": 4859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000023.wav"
+            ]
+        }
+    },
+    {
+        "id": 4862,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003077.wav"
+            ]
+        }
+    },
+    {
+        "id": 4872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008788.wav"
+            ]
+        }
+    },
+    {
+        "id": 4876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005325.wav"
+            ]
+        }
+    },
+    {
+        "id": 4888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002649.wav"
+            ]
+        }
+    },
+    {
+        "id": 4893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000438.wav"
+            ]
+        }
+    },
+    {
+        "id": 4899,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008011.wav"
+            ]
+        }
+    },
+    {
+        "id": 4913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007293.wav"
+            ]
+        }
+    },
+    {
+        "id": 4921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001001.wav"
+            ]
+        }
+    },
+    {
+        "id": 4932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001911.wav"
+            ]
+        }
+    },
+    {
+        "id": 4935,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003168.wav"
+            ]
+        }
+    },
+    {
+        "id": 4942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003506.wav"
+            ]
+        }
+    },
+    {
+        "id": 4946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008918.wav"
+            ]
+        }
+    },
+    {
+        "id": 4959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005542.wav"
+            ]
+        }
+    },
+    {
+        "id": 4964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000043.wav"
+            ]
+        }
+    },
+    {
+        "id": 4976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005283.wav"
+            ]
+        }
+    },
+    {
+        "id": 4979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003658.wav"
+            ]
+        }
+    },
+    {
+        "id": 4985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001708.wav"
+            ]
+        }
+    },
+    {
+        "id": 4987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet playing longer than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005199.wav"
+            ]
+        }
+    },
+    {
+        "id": 4988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007707.wav"
+            ]
+        }
+    },
+    {
+        "id": 5004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003050.wav"
+            ]
+        }
+    },
+    {
+        "id": 5006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008843.wav"
+            ]
+        }
+    },
+    {
+        "id": 5007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007558.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007558.wav"
+            ]
+        }
+    },
+    {
+        "id": 5009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002236.wav"
+            ]
+        }
+    },
+    {
+        "id": 5018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006354.wav"
+            ]
+        }
+    },
+    {
+        "id": 5022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004483.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004483.wav"
+            ]
+        }
+    },
+    {
+        "id": 5025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002763.wav"
+            ]
+        }
+    },
+    {
+        "id": 5026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008157.wav"
+            ]
+        }
+    },
+    {
+        "id": 5029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003936.wav"
+            ]
+        }
+    },
+    {
+        "id": 5034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004704.wav"
+            ]
+        }
+    },
+    {
+        "id": 5037,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001535.wav"
+            ]
+        }
+    },
+    {
+        "id": 5039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003406.wav"
+            ]
+        }
+    },
+    {
+        "id": 5043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008734.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008734.wav"
+            ]
+        }
+    },
+    {
+        "id": 5046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007625.wav"
+            ]
+        }
+    },
+    {
+        "id": 5049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007777.wav"
+            ]
+        }
+    },
+    {
+        "id": 5050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005659.wav"
+            ]
+        }
+    },
+    {
+        "id": 5053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007875.wav"
+            ]
+        }
+    },
+    {
+        "id": 5054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003164.wav"
+            ]
+        }
+    },
+    {
+        "id": 5059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001370.wav"
+            ]
+        }
+    },
+    {
+        "id": 5063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002176.wav"
+            ]
+        }
+    },
+    {
+        "id": 5069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006687.wav"
+            ]
+        }
+    },
+    {
+        "id": 5095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003226.wav"
+            ]
+        }
+    },
+    {
+        "id": 5098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004332.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004332.wav"
+            ]
+        }
+    },
+    {
+        "id": 5100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007132.wav"
+            ]
+        }
+    },
+    {
+        "id": 5102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001588.wav"
+            ]
+        }
+    },
+    {
+        "id": 5105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006952.wav"
+            ]
+        }
+    },
+    {
+        "id": 5109,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002708.wav"
+            ]
+        }
+    },
+    {
+        "id": 5115,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003857.wav"
+            ]
+        }
+    },
+    {
+        "id": 5146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000352.wav"
+            ]
+        }
+    },
+    {
+        "id": 5159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004080.wav"
+            ]
+        }
+    },
+    {
+        "id": 5172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000528.wav"
+            ]
+        }
+    },
+    {
+        "id": 5175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006988.wav"
+            ]
+        }
+    },
+    {
+        "id": 5181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001325.wav"
+            ]
+        }
+    },
+    {
+        "id": 5184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003855.wav"
+            ]
+        }
+    },
+    {
+        "id": 5191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007236.wav"
+            ]
+        }
+    },
+    {
+        "id": 5204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002337.wav"
+            ]
+        }
+    },
+    {
+        "id": 5206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000220.wav"
+            ]
+        }
+    },
+    {
+        "id": 5209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007313.wav"
+            ]
+        }
+    },
+    {
+        "id": 5211,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004865.wav"
+            ]
+        }
+    },
+    {
+        "id": 5216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007766.wav"
+            ]
+        }
+    },
+    {
+        "id": 5218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004141.wav"
+            ]
+        }
+    },
+    {
+        "id": 5222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001730.wav"
+            ]
+        }
+    },
+    {
+        "id": 5233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004568.wav"
+            ]
+        }
+    },
+    {
+        "id": 5249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008214.wav"
+            ]
+        }
+    },
+    {
+        "id": 5250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005119.wav"
+            ]
+        }
+    },
+    {
+        "id": 5251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001197.wav"
+            ]
+        }
+    },
+    {
+        "id": 5275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008796.wav"
+            ]
+        }
+    },
+    {
+        "id": 5282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003199.wav"
+            ]
+        }
+    },
+    {
+        "id": 5287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008396.wav"
+            ]
+        }
+    },
+    {
+        "id": 5290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003611.wav"
+            ]
+        }
+    },
+    {
+        "id": 5291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005620.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005620.wav"
+            ]
+        }
+    },
+    {
+        "id": 5296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007431.wav"
+            ]
+        }
+    },
+    {
+        "id": 5297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002825.wav"
+            ]
+        }
+    },
+    {
+        "id": 5302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005014.wav"
+            ]
+        }
+    },
+    {
+        "id": 5304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002139.wav"
+            ]
+        }
+    },
+    {
+        "id": 5306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008941.wav"
+            ]
+        }
+    },
+    {
+        "id": 5317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe on the middle more rhythmic than the electric_bass on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007348.wav"
+            ]
+        }
+    },
+    {
+        "id": 5321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007030.wav"
+            ]
+        }
+    },
+    {
+        "id": 5333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003230.wav"
+            ]
+        }
+    },
+    {
+        "id": 5345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004012.wav"
+            ]
+        }
+    },
+    {
+        "id": 5354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004929.wav"
+            ]
+        }
+    },
+    {
+        "id": 5355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006494.wav"
+            ]
+        }
+    },
+    {
+        "id": 5356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004667.wav"
+            ]
+        }
+    },
+    {
+        "id": 5360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001252.wav"
+            ]
+        }
+    },
+    {
+        "id": 5362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004407.wav"
+            ]
+        }
+    },
+    {
+        "id": 5363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the middle more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003423.wav"
+            ]
+        }
+    },
+    {
+        "id": 5367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007929.wav"
+            ]
+        }
+    },
+    {
+        "id": 5370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001972.wav"
+            ]
+        }
+    },
+    {
+        "id": 5374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002945.wav"
+            ]
+        }
+    },
+    {
+        "id": 5377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000543.wav"
+            ]
+        }
+    },
+    {
+        "id": 5381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the right more rhythmic than the banjo on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005225.wav"
+            ]
+        }
+    },
+    {
+        "id": 5389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000877.wav"
+            ]
+        }
+    },
+    {
+        "id": 5394,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000705.wav"
+            ]
+        }
+    },
+    {
+        "id": 5397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003181.wav"
+            ]
+        }
+    },
+    {
+        "id": 5401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006491.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006491.wav"
+            ]
+        }
+    },
+    {
+        "id": 5406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004029.wav"
+            ]
+        }
+    },
+    {
+        "id": 5409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006669.wav"
+            ]
+        }
+    },
+    {
+        "id": 5436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008364.wav"
+            ]
+        }
+    },
+    {
+        "id": 5452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004480.wav"
+            ]
+        }
+    },
+    {
+        "id": 5453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas on the left louder than the congas on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004368.wav"
+            ]
+        }
+    },
+    {
+        "id": 5461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000781.wav"
+            ]
+        }
+    },
+    {
+        "id": 5474,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006977.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006977.wav"
+            ]
+        }
+    },
+    {
+        "id": 5478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003687.wav"
+            ]
+        }
+    },
+    {
+        "id": 5481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000612.wav"
+            ]
+        }
+    },
+    {
+        "id": 5482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003485.wav"
+            ]
+        }
+    },
+    {
+        "id": 5486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007737.wav"
+            ]
+        }
+    },
+    {
+        "id": 5487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007579.wav"
+            ]
+        }
+    },
+    {
+        "id": 5497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006146.wav"
+            ]
+        }
+    },
+    {
+        "id": 5505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006449.wav"
+            ]
+        }
+    },
+    {
+        "id": 5527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000489.wav"
+            ]
+        }
+    },
+    {
+        "id": 5532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003033.wav"
+            ]
+        }
+    },
+    {
+        "id": 5533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007045.wav"
+            ]
+        }
+    },
+    {
+        "id": 5538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008396.wav"
+            ]
+        }
+    },
+    {
+        "id": 5543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000798.wav"
+            ]
+        }
+    },
+    {
+        "id": 5544,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007737.wav"
+            ]
+        }
+    },
+    {
+        "id": 5546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003505.wav"
+            ]
+        }
+    },
+    {
+        "id": 5560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002996.wav"
+            ]
+        }
+    },
+    {
+        "id": 5564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006067.wav"
+            ]
+        }
+    },
+    {
+        "id": 5565,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007787.wav"
+            ]
+        }
+    },
+    {
+        "id": 5567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003550.wav"
+            ]
+        }
+    },
+    {
+        "id": 5575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001949.wav"
+            ]
+        }
+    },
+    {
+        "id": 5580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001312.wav"
+            ]
+        }
+    },
+    {
+        "id": 5582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002237.wav"
+            ]
+        }
+    },
+    {
+        "id": 5588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001115.wav"
+            ]
+        }
+    },
+    {
+        "id": 5595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006353.wav"
+            ]
+        }
+    },
+    {
+        "id": 5596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008340.wav"
+            ]
+        }
+    },
+    {
+        "id": 5597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005884.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005884.wav"
+            ]
+        }
+    },
+    {
+        "id": 5600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006805.wav"
+            ]
+        }
+    },
+    {
+        "id": 5607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002611.wav"
+            ]
+        }
+    },
+    {
+        "id": 5609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002808.wav"
+            ]
+        }
+    },
+    {
+        "id": 5614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004064.wav"
+            ]
+        }
+    },
+    {
+        "id": 5619,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000226.wav"
+            ]
+        }
+    },
+    {
+        "id": 5622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004275.wav"
+            ]
+        }
+    },
+    {
+        "id": 5628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002475.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002475.wav"
+            ]
+        }
+    },
+    {
+        "id": 5635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003213.wav"
+            ]
+        }
+    },
+    {
+        "id": 5641,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004668.wav"
+            ]
+        }
+    },
+    {
+        "id": 5646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004104.wav"
+            ]
+        }
+    },
+    {
+        "id": 5651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003861.wav"
+            ]
+        }
+    },
+    {
+        "id": 5652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003302.wav"
+            ]
+        }
+    },
+    {
+        "id": 5655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004912.wav"
+            ]
+        }
+    },
+    {
+        "id": 5656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002994.wav"
+            ]
+        }
+    },
+    {
+        "id": 5658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004765.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004765.wav"
+            ]
+        }
+    },
+    {
+        "id": 5659,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004882.wav"
+            ]
+        }
+    },
+    {
+        "id": 5663,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000326.wav"
+            ]
+        }
+    },
+    {
+        "id": 5665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001877.wav"
+            ]
+        }
+    },
+    {
+        "id": 5668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002485.wav"
+            ]
+        }
+    },
+    {
+        "id": 5672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002397.wav"
+            ]
+        }
+    },
+    {
+        "id": 5673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008027.wav"
+            ]
+        }
+    },
+    {
+        "id": 5680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008228.wav"
+            ]
+        }
+    },
+    {
+        "id": 5683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007141.wav"
+            ]
+        }
+    },
+    {
+        "id": 5686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001014.wav"
+            ]
+        }
+    },
+    {
+        "id": 5687,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008879.wav"
+            ]
+        }
+    },
+    {
+        "id": 5699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005562.wav"
+            ]
+        }
+    },
+    {
+        "id": 5702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001723.wav"
+            ]
+        }
+    },
+    {
+        "id": 5713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002159.wav"
+            ]
+        }
+    },
+    {
+        "id": 5716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007005.wav"
+            ]
+        }
+    },
+    {
+        "id": 5717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005246.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005246.wav"
+            ]
+        }
+    },
+    {
+        "id": 5718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002584.wav"
+            ]
+        }
+    },
+    {
+        "id": 5739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006166.wav"
+            ]
+        }
+    },
+    {
+        "id": 5768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000002.wav"
+            ]
+        }
+    },
+    {
+        "id": 5780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001476.wav"
+            ]
+        }
+    },
+    {
+        "id": 5788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000961.wav"
+            ]
+        }
+    },
+    {
+        "id": 5791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001143.wav"
+            ]
+        }
+    },
+    {
+        "id": 5803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001015.wav"
+            ]
+        }
+    },
+    {
+        "id": 5805,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002220.wav"
+            ]
+        }
+    },
+    {
+        "id": 5811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005612.wav"
+            ]
+        }
+    },
+    {
+        "id": 5814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001152.wav"
+            ]
+        }
+    },
+    {
+        "id": 5819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000951.wav"
+            ]
+        }
+    },
+    {
+        "id": 5820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006828.wav"
+            ]
+        }
+    },
+    {
+        "id": 5822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006145.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006145.wav"
+            ]
+        }
+    },
+    {
+        "id": 5827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002672.wav"
+            ]
+        }
+    },
+    {
+        "id": 5829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003757.wav"
+            ]
+        }
+    },
+    {
+        "id": 5830,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004841.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004841.wav"
+            ]
+        }
+    },
+    {
+        "id": 5833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001700.wav"
+            ]
+        }
+    },
+    {
+        "id": 5844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004396.wav"
+            ]
+        }
+    },
+    {
+        "id": 5853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005861.wav"
+            ]
+        }
+    },
+    {
+        "id": 5856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003179.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003179.wav"
+            ]
+        }
+    },
+    {
+        "id": 5858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007812.wav"
+            ]
+        }
+    },
+    {
+        "id": 5859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002932.wav"
+            ]
+        }
+    },
+    {
+        "id": 5860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003479.wav"
+            ]
+        }
+    },
+    {
+        "id": 5864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001474.wav"
+            ]
+        }
+    },
+    {
+        "id": 5865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000784.wav"
+            ]
+        }
+    },
+    {
+        "id": 5866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003143.wav"
+            ]
+        }
+    },
+    {
+        "id": 5872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006654.wav"
+            ]
+        }
+    },
+    {
+        "id": 5877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008701.wav"
+            ]
+        }
+    },
+    {
+        "id": 5880,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004230.wav"
+            ]
+        }
+    },
+    {
+        "id": 5888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008926.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008926.wav"
+            ]
+        }
+    },
+    {
+        "id": 5892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002294.wav"
+            ]
+        }
+    },
+    {
+        "id": 5905,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008668.wav"
+            ]
+        }
+    },
+    {
+        "id": 5914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005176.wav"
+            ]
+        }
+    },
+    {
+        "id": 5926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003321.wav"
+            ]
+        }
+    },
+    {
+        "id": 5938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005417.wav"
+            ]
+        }
+    },
+    {
+        "id": 5942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005431.wav"
+            ]
+        }
+    },
+    {
+        "id": 5952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008564.wav"
+            ]
+        }
+    },
+    {
+        "id": 5953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008668.wav"
+            ]
+        }
+    },
+    {
+        "id": 5954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006559.wav"
+            ]
+        }
+    },
+    {
+        "id": 5957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003618.wav"
+            ]
+        }
+    },
+    {
+        "id": 5961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001038.wav"
+            ]
+        }
+    },
+    {
+        "id": 5974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007656.wav"
+            ]
+        }
+    },
+    {
+        "id": 5991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003633.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003633.wav"
+            ]
+        }
+    },
+    {
+        "id": 5995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006191.wav"
+            ]
+        }
+    },
+    {
+        "id": 5999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002809.wav"
+            ]
+        }
+    },
+    {
+        "id": 6005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007483.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007483.wav"
+            ]
+        }
+    },
+    {
+        "id": 6007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008668.wav"
+            ]
+        }
+    },
+    {
+        "id": 6012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003007.wav"
+            ]
+        }
+    },
+    {
+        "id": 6013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008680.wav"
+            ]
+        }
+    },
+    {
+        "id": 6016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007913.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007913.wav"
+            ]
+        }
+    },
+    {
+        "id": 6020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008364.wav"
+            ]
+        }
+    },
+    {
+        "id": 6021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005747.wav"
+            ]
+        }
+    },
+    {
+        "id": 6023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007210.wav"
+            ]
+        }
+    },
+    {
+        "id": 6027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002696.wav"
+            ]
+        }
+    },
+    {
+        "id": 6031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000243.wav"
+            ]
+        }
+    },
+    {
+        "id": 6040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000604.wav"
+            ]
+        }
+    },
+    {
+        "id": 6049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003920.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003920.wav"
+            ]
+        }
+    },
+    {
+        "id": 6052,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006630.wav"
+            ]
+        }
+    },
+    {
+        "id": 6070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005856.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005856.wav"
+            ]
+        }
+    },
+    {
+        "id": 6072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008606.wav"
+            ]
+        }
+    },
+    {
+        "id": 6073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006848.wav"
+            ]
+        }
+    },
+    {
+        "id": 6076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001373.wav"
+            ]
+        }
+    },
+    {
+        "id": 6085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000075.wav"
+            ]
+        }
+    },
+    {
+        "id": 6090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004775.wav"
+            ]
+        }
+    },
+    {
+        "id": 6096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006729.wav"
+            ]
+        }
+    },
+    {
+        "id": 6107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002635.wav"
+            ]
+        }
+    },
+    {
+        "id": 6108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000640.wav"
+            ]
+        }
+    },
+    {
+        "id": 6116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001071.wav"
+            ]
+        }
+    },
+    {
+        "id": 6123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left more rhythmic than the acoustic_guitar on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004630.wav"
+            ]
+        }
+    },
+    {
+        "id": 6127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002370.wav"
+            ]
+        }
+    },
+    {
+        "id": 6133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008861.wav"
+            ]
+        }
+    },
+    {
+        "id": 6135,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005106.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005106.wav"
+            ]
+        }
+    },
+    {
+        "id": 6136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003148.wav"
+            ]
+        }
+    },
+    {
+        "id": 6148,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005832.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005832.wav"
+            ]
+        }
+    },
+    {
+        "id": 6150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007397.wav"
+            ]
+        }
+    },
+    {
+        "id": 6151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005095.wav"
+            ]
+        }
+    },
+    {
+        "id": 6182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000831.wav"
+            ]
+        }
+    },
+    {
+        "id": 6187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002355.wav"
+            ]
+        }
+    },
+    {
+        "id": 6197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007137.wav"
+            ]
+        }
+    },
+    {
+        "id": 6202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001962.wav"
+            ]
+        }
+    },
+    {
+        "id": 6209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004196.wav"
+            ]
+        }
+    },
+    {
+        "id": 6225,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005092.wav"
+            ]
+        }
+    },
+    {
+        "id": 6243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008831.wav"
+            ]
+        }
+    },
+    {
+        "id": 6246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas on the right more rhythmic than the xylophone on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005671.wav"
+            ]
+        }
+    },
+    {
+        "id": 6247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004457.wav"
+            ]
+        }
+    },
+    {
+        "id": 6260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006303.wav"
+            ]
+        }
+    },
+    {
+        "id": 6266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003707.wav"
+            ]
+        }
+    },
+    {
+        "id": 6272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001404.wav"
+            ]
+        }
+    },
+    {
+        "id": 6278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004423.wav"
+            ]
+        }
+    },
+    {
+        "id": 6281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001569.wav"
+            ]
+        }
+    },
+    {
+        "id": 6284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005020.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005020.wav"
+            ]
+        }
+    },
+    {
+        "id": 6287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008424.wav"
+            ]
+        }
+    },
+    {
+        "id": 6293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004718.wav"
+            ]
+        }
+    },
+    {
+        "id": 6295,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007820.wav"
+            ]
+        }
+    },
+    {
+        "id": 6303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003803.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003803.wav"
+            ]
+        }
+    },
+    {
+        "id": 6306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008103.wav"
+            ]
+        }
+    },
+    {
+        "id": 6308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001024.wav"
+            ]
+        }
+    },
+    {
+        "id": 6313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001197.wav"
+            ]
+        }
+    },
+    {
+        "id": 6321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003463.wav"
+            ]
+        }
+    },
+    {
+        "id": 6324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006470.wav"
+            ]
+        }
+    },
+    {
+        "id": 6327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008481.wav"
+            ]
+        }
+    },
+    {
+        "id": 6339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008162.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008162.wav"
+            ]
+        }
+    },
+    {
+        "id": 6343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003434.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003434.wav"
+            ]
+        }
+    },
+    {
+        "id": 6365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000243.wav"
+            ]
+        }
+    },
+    {
+        "id": 6368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004886.wav"
+            ]
+        }
+    },
+    {
+        "id": 6371,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001536.wav"
+            ]
+        }
+    },
+    {
+        "id": 6387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002205.wav"
+            ]
+        }
+    },
+    {
+        "id": 6392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000168.wav"
+            ]
+        }
+    },
+    {
+        "id": 6397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the fifth instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004113.wav"
+            ]
+        }
+    },
+    {
+        "id": 6400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006217.wav"
+            ]
+        }
+    },
+    {
+        "id": 6408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003098.wav"
+            ]
+        }
+    },
+    {
+        "id": 6410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008361.wav"
+            ]
+        }
+    },
+    {
+        "id": 6413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002115.wav"
+            ]
+        }
+    },
+    {
+        "id": 6419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007980.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007980.wav"
+            ]
+        }
+    },
+    {
+        "id": 6421,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008813.wav"
+            ]
+        }
+    },
+    {
+        "id": 6422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001523.wav"
+            ]
+        }
+    },
+    {
+        "id": 6423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000973.wav"
+            ]
+        }
+    },
+    {
+        "id": 6429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001442.wav"
+            ]
+        }
+    },
+    {
+        "id": 6432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004998.wav"
+            ]
+        }
+    },
+    {
+        "id": 6435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006481.wav"
+            ]
+        }
+    },
+    {
+        "id": 6438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008808.wav"
+            ]
+        }
+    },
+    {
+        "id": 6443,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008912.wav"
+            ]
+        }
+    },
+    {
+        "id": 6447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000696.wav"
+            ]
+        }
+    },
+    {
+        "id": 6448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003338.wav"
+            ]
+        }
+    },
+    {
+        "id": 6457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000054.wav"
+            ]
+        }
+    },
+    {
+        "id": 6467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the middle louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004437.wav"
+            ]
+        }
+    },
+    {
+        "id": 6468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005636.wav"
+            ]
+        }
+    },
+    {
+        "id": 6469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006074.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006074.wav"
+            ]
+        }
+    },
+    {
+        "id": 6473,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003254.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003254.wav"
+            ]
+        }
+    },
+    {
+        "id": 6476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000474.wav"
+            ]
+        }
+    },
+    {
+        "id": 6478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002304.wav"
+            ]
+        }
+    },
+    {
+        "id": 6480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007182.wav"
+            ]
+        }
+    },
+    {
+        "id": 6489,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002470.wav"
+            ]
+        }
+    },
+    {
+        "id": 6494,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007806.wav"
+            ]
+        }
+    },
+    {
+        "id": 6495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the fourth instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003605.wav"
+            ]
+        }
+    },
+    {
+        "id": 6497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007632.wav"
+            ]
+        }
+    },
+    {
+        "id": 6500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008207.wav"
+            ]
+        }
+    },
+    {
+        "id": 6503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001094.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001094.wav"
+            ]
+        }
+    },
+    {
+        "id": 6504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003316.wav"
+            ]
+        }
+    },
+    {
+        "id": 6517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003230.wav"
+            ]
+        }
+    },
+    {
+        "id": 6521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008844.wav"
+            ]
+        }
+    },
+    {
+        "id": 6543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005152.wav"
+            ]
+        }
+    },
+    {
+        "id": 6551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008147.wav"
+            ]
+        }
+    },
+    {
+        "id": 6556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003280.wav"
+            ]
+        }
+    },
+    {
+        "id": 6560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005368.wav"
+            ]
+        }
+    },
+    {
+        "id": 6561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004958.wav"
+            ]
+        }
+    },
+    {
+        "id": 6563,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008232.wav"
+            ]
+        }
+    },
+    {
+        "id": 6566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002154.wav"
+            ]
+        }
+    },
+    {
+        "id": 6573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008932.wav"
+            ]
+        }
+    },
+    {
+        "id": 6578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005142.wav"
+            ]
+        }
+    },
+    {
+        "id": 6585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003863.wav"
+            ]
+        }
+    },
+    {
+        "id": 6587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007515.wav"
+            ]
+        }
+    },
+    {
+        "id": 6590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007022.wav"
+            ]
+        }
+    },
+    {
+        "id": 6591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006966.wav"
+            ]
+        }
+    },
+    {
+        "id": 6615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008507.wav"
+            ]
+        }
+    },
+    {
+        "id": 6618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003782.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003782.wav"
+            ]
+        }
+    },
+    {
+        "id": 6620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001400.wav"
+            ]
+        }
+    },
+    {
+        "id": 6622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007661.wav"
+            ]
+        }
+    },
+    {
+        "id": 6623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001515.wav"
+            ]
+        }
+    },
+    {
+        "id": 6641,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002847.wav"
+            ]
+        }
+    },
+    {
+        "id": 6651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000340.wav"
+            ]
+        }
+    },
+    {
+        "id": 6657,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004350.wav"
+            ]
+        }
+    },
+    {
+        "id": 6675,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008471.wav"
+            ]
+        }
+    },
+    {
+        "id": 6678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007096.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007096.wav"
+            ]
+        }
+    },
+    {
+        "id": 6681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008124.wav"
+            ]
+        }
+    },
+    {
+        "id": 6686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002420.wav"
+            ]
+        }
+    },
+    {
+        "id": 6697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007760.wav"
+            ]
+        }
+    },
+    {
+        "id": 6703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008044.wav"
+            ]
+        }
+    },
+    {
+        "id": 6709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005207.wav"
+            ]
+        }
+    },
+    {
+        "id": 6712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003449.wav"
+            ]
+        }
+    },
+    {
+        "id": 6721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002242.wav"
+            ]
+        }
+    },
+    {
+        "id": 6724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007169.wav"
+            ]
+        }
+    },
+    {
+        "id": 6725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004635.wav"
+            ]
+        }
+    },
+    {
+        "id": 6727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003257.wav"
+            ]
+        }
+    },
+    {
+        "id": 6733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006121.wav"
+            ]
+        }
+    },
+    {
+        "id": 6734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002119.wav"
+            ]
+        }
+    },
+    {
+        "id": 6738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000170.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000170.wav"
+            ]
+        }
+    },
+    {
+        "id": 6751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004386.wav"
+            ]
+        }
+    },
+    {
+        "id": 6760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000786.wav"
+            ]
+        }
+    },
+    {
+        "id": 6768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005327.wav"
+            ]
+        }
+    },
+    {
+        "id": 6769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008267.wav"
+            ]
+        }
+    },
+    {
+        "id": 6780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004416.wav"
+            ]
+        }
+    },
+    {
+        "id": 6788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007820.wav"
+            ]
+        }
+    },
+    {
+        "id": 6792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004746.wav"
+            ]
+        }
+    },
+    {
+        "id": 6794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006656.wav"
+            ]
+        }
+    },
+    {
+        "id": 6825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000761.wav"
+            ]
+        }
+    },
+    {
+        "id": 6829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006648.wav"
+            ]
+        }
+    },
+    {
+        "id": 6833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008501.wav"
+            ]
+        }
+    },
+    {
+        "id": 6834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004573.wav"
+            ]
+        }
+    },
+    {
+        "id": 6835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003789.wav"
+            ]
+        }
+    },
+    {
+        "id": 6845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003354.wav"
+            ]
+        }
+    },
+    {
+        "id": 6848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006213.wav"
+            ]
+        }
+    },
+    {
+        "id": 6853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000083.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000083.wav"
+            ]
+        }
+    },
+    {
+        "id": 6854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007290.wav"
+            ]
+        }
+    },
+    {
+        "id": 6865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001519.wav"
+            ]
+        }
+    },
+    {
+        "id": 6866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002161.wav"
+            ]
+        }
+    },
+    {
+        "id": 6870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008886.wav"
+            ]
+        }
+    },
+    {
+        "id": 6872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006223.wav"
+            ]
+        }
+    },
+    {
+        "id": 6876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004084.wav"
+            ]
+        }
+    },
+    {
+        "id": 6887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002900.wav"
+            ]
+        }
+    },
+    {
+        "id": 6888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005543.wav"
+            ]
+        }
+    },
+    {
+        "id": 6889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000246.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000246.wav"
+            ]
+        }
+    },
+    {
+        "id": 6892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001115.wav"
+            ]
+        }
+    },
+    {
+        "id": 6894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004359.wav"
+            ]
+        }
+    },
+    {
+        "id": 6900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000418.wav"
+            ]
+        }
+    },
+    {
+        "id": 6909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008637.wav"
+            ]
+        }
+    },
+    {
+        "id": 6911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006865.wav"
+            ]
+        }
+    },
+    {
+        "id": 6912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002869.wav"
+            ]
+        }
+    },
+    {
+        "id": 6918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003925.wav"
+            ]
+        }
+    },
+    {
+        "id": 6922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002206.wav"
+            ]
+        }
+    },
+    {
+        "id": 6923,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004644.wav"
+            ]
+        }
+    },
+    {
+        "id": 6924,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the violin on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000329.wav"
+            ]
+        }
+    },
+    {
+        "id": 6925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000479.wav"
+            ]
+        }
+    },
+    {
+        "id": 6937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004070.wav"
+            ]
+        }
+    },
+    {
+        "id": 6941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002910.wav"
+            ]
+        }
+    },
+    {
+        "id": 6973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008526.wav"
+            ]
+        }
+    },
+    {
+        "id": 6981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008573.wav"
+            ]
+        }
+    },
+    {
+        "id": 6983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002033.wav"
+            ]
+        }
+    },
+    {
+        "id": 6992,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008418.wav"
+            ]
+        }
+    },
+    {
+        "id": 6996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002979.wav"
+            ]
+        }
+    },
+    {
+        "id": 7006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002836.wav"
+            ]
+        }
+    },
+    {
+        "id": 7009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006618.wav"
+            ]
+        }
+    },
+    {
+        "id": 7010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000334.wav"
+            ]
+        }
+    },
+    {
+        "id": 7016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006163.wav"
+            ]
+        }
+    },
+    {
+        "id": 7025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000362.wav"
+            ]
+        }
+    },
+    {
+        "id": 7030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000963.wav"
+            ]
+        }
+    },
+    {
+        "id": 7031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001042.wav"
+            ]
+        }
+    },
+    {
+        "id": 7032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006995.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006995.wav"
+            ]
+        }
+    },
+    {
+        "id": 7037,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001538.wav"
+            ]
+        }
+    },
+    {
+        "id": 7054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003535.wav"
+            ]
+        }
+    },
+    {
+        "id": 7059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003596.wav"
+            ]
+        }
+    },
+    {
+        "id": 7079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right more rhythmic than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004183.wav"
+            ]
+        }
+    },
+    {
+        "id": 7088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003408.wav"
+            ]
+        }
+    },
+    {
+        "id": 7102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008949.wav"
+            ]
+        }
+    },
+    {
+        "id": 7104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004127.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004127.wav"
+            ]
+        }
+    },
+    {
+        "id": 7107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000950.wav"
+            ]
+        }
+    },
+    {
+        "id": 7112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004563.wav"
+            ]
+        }
+    },
+    {
+        "id": 7120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004807.wav"
+            ]
+        }
+    },
+    {
+        "id": 7132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007348.wav"
+            ]
+        }
+    },
+    {
+        "id": 7133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007121.wav"
+            ]
+        }
+    },
+    {
+        "id": 7136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003470.wav"
+            ]
+        }
+    },
+    {
+        "id": 7137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007731.wav"
+            ]
+        }
+    },
+    {
+        "id": 7138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008750.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008750.wav"
+            ]
+        }
+    },
+    {
+        "id": 7139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005824.wav"
+            ]
+        }
+    },
+    {
+        "id": 7143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008828.wav"
+            ]
+        }
+    },
+    {
+        "id": 7151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004907.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004907.wav"
+            ]
+        }
+    },
+    {
+        "id": 7162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007423.wav"
+            ]
+        }
+    },
+    {
+        "id": 7164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005245.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005245.wav"
+            ]
+        }
+    },
+    {
+        "id": 7182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007616.wav"
+            ]
+        }
+    },
+    {
+        "id": 7184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005640.wav"
+            ]
+        }
+    },
+    {
+        "id": 7185,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008105.wav"
+            ]
+        }
+    },
+    {
+        "id": 7195,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006600.wav"
+            ]
+        }
+    },
+    {
+        "id": 7198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008446.wav"
+            ]
+        }
+    },
+    {
+        "id": 7199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the clarinet on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006461.wav"
+            ]
+        }
+    },
+    {
+        "id": 7200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002662.wav"
+            ]
+        }
+    },
+    {
+        "id": 7209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004825.wav"
+            ]
+        }
+    },
+    {
+        "id": 7213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding banjo in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004619.wav"
+            ]
+        }
+    },
+    {
+        "id": 7222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005876.wav"
+            ]
+        }
+    },
+    {
+        "id": 7228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000710.wav"
+            ]
+        }
+    },
+    {
+        "id": 7229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008901.wav"
+            ]
+        }
+    },
+    {
+        "id": 7231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001385.wav"
+            ]
+        }
+    },
+    {
+        "id": 7237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002359.wav"
+            ]
+        }
+    },
+    {
+        "id": 7245,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007915.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007915.wav"
+            ]
+        }
+    },
+    {
+        "id": 7252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001313.wav"
+            ]
+        }
+    },
+    {
+        "id": 7253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002013.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002013.wav"
+            ]
+        }
+    },
+    {
+        "id": 7260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000852.wav"
+            ]
+        }
+    },
+    {
+        "id": 7267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000028.wav"
+            ]
+        }
+    },
+    {
+        "id": 7268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006929.wav"
+            ]
+        }
+    },
+    {
+        "id": 7270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003410.wav"
+            ]
+        }
+    },
+    {
+        "id": 7278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004766.wav"
+            ]
+        }
+    },
+    {
+        "id": 7280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000806.wav"
+            ]
+        }
+    },
+    {
+        "id": 7284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007191.wav"
+            ]
+        }
+    },
+    {
+        "id": 7293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008318.wav"
+            ]
+        }
+    },
+    {
+        "id": 7294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000565.wav"
+            ]
+        }
+    },
+    {
+        "id": 7297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004128.wav"
+            ]
+        }
+    },
+    {
+        "id": 7304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000021.wav"
+            ]
+        }
+    },
+    {
+        "id": 7313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006047.wav"
+            ]
+        }
+    },
+    {
+        "id": 7327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008111.wav"
+            ]
+        }
+    },
+    {
+        "id": 7334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002606.wav"
+            ]
+        }
+    },
+    {
+        "id": 7342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001716.wav"
+            ]
+        }
+    },
+    {
+        "id": 7343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008370.wav"
+            ]
+        }
+    },
+    {
+        "id": 7347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006907.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006907.wav"
+            ]
+        }
+    },
+    {
+        "id": 7352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001261.wav"
+            ]
+        }
+    },
+    {
+        "id": 7354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001516.wav"
+            ]
+        }
+    },
+    {
+        "id": 7356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003561.wav"
+            ]
+        }
+    },
+    {
+        "id": 7357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the middle louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000680.wav"
+            ]
+        }
+    },
+    {
+        "id": 7358,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005604.wav"
+            ]
+        }
+    },
+    {
+        "id": 7362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006800.wav"
+            ]
+        }
+    },
+    {
+        "id": 7367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005485.wav"
+            ]
+        }
+    },
+    {
+        "id": 7369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006621.wav"
+            ]
+        }
+    },
+    {
+        "id": 7384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003250.wav"
+            ]
+        }
+    },
+    {
+        "id": 7386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002664.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002664.wav"
+            ]
+        }
+    },
+    {
+        "id": 7412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005973.wav"
+            ]
+        }
+    },
+    {
+        "id": 7415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000189.wav"
+            ]
+        }
+    },
+    {
+        "id": 7422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right more rhythmic than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002109.wav"
+            ]
+        }
+    },
+    {
+        "id": 7429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005740.wav"
+            ]
+        }
+    },
+    {
+        "id": 7457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000175.wav"
+            ]
+        }
+    },
+    {
+        "id": 7468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005530.wav"
+            ]
+        }
+    },
+    {
+        "id": 7476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005311.wav"
+            ]
+        }
+    },
+    {
+        "id": 7478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004124.wav"
+            ]
+        }
+    },
+    {
+        "id": 7482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000963.wav"
+            ]
+        }
+    },
+    {
+        "id": 7487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006505.wav"
+            ]
+        }
+    },
+    {
+        "id": 7496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005967.wav"
+            ]
+        }
+    },
+    {
+        "id": 7500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008109.wav"
+            ]
+        }
+    },
+    {
+        "id": 7503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000897.wav"
+            ]
+        }
+    },
+    {
+        "id": 7509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003040.wav"
+            ]
+        }
+    },
+    {
+        "id": 7517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004707.wav"
+            ]
+        }
+    },
+    {
+        "id": 7518,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007502.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007502.wav"
+            ]
+        }
+    },
+    {
+        "id": 7521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000186.wav"
+            ]
+        }
+    },
+    {
+        "id": 7537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007760.wav"
+            ]
+        }
+    },
+    {
+        "id": 7540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008830.wav"
+            ]
+        }
+    },
+    {
+        "id": 7561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002596.wav"
+            ]
+        }
+    },
+    {
+        "id": 7564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004178.wav"
+            ]
+        }
+    },
+    {
+        "id": 7567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003611.wav"
+            ]
+        }
+    },
+    {
+        "id": 7577,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003225.wav"
+            ]
+        }
+    },
+    {
+        "id": 7584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003842.wav"
+            ]
+        }
+    },
+    {
+        "id": 7598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004900.wav"
+            ]
+        }
+    },
+    {
+        "id": 7603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000413.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000413.wav"
+            ]
+        }
+    },
+    {
+        "id": 7616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002237.wav"
+            ]
+        }
+    },
+    {
+        "id": 7620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002298.wav"
+            ]
+        }
+    },
+    {
+        "id": 7621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003042.wav"
+            ]
+        }
+    },
+    {
+        "id": 7622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008203.wav"
+            ]
+        }
+    },
+    {
+        "id": 7624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008392.wav"
+            ]
+        }
+    },
+    {
+        "id": 7627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006850.wav"
+            ]
+        }
+    },
+    {
+        "id": 7628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000234.wav"
+            ]
+        }
+    },
+    {
+        "id": 7631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008403.wav"
+            ]
+        }
+    },
+    {
+        "id": 7646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000169.wav"
+            ]
+        }
+    },
+    {
+        "id": 7649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000583.wav"
+            ]
+        }
+    },
+    {
+        "id": 7652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004448.wav"
+            ]
+        }
+    },
+    {
+        "id": 7668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000360.wav"
+            ]
+        }
+    },
+    {
+        "id": 7684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003450.wav"
+            ]
+        }
+    },
+    {
+        "id": 7697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002974.wav"
+            ]
+        }
+    },
+    {
+        "id": 7698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005672.wav"
+            ]
+        }
+    },
+    {
+        "id": 7702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001080.wav"
+            ]
+        }
+    },
+    {
+        "id": 7703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003218.wav"
+            ]
+        }
+    },
+    {
+        "id": 7704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005196.wav"
+            ]
+        }
+    },
+    {
+        "id": 7707,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003805.wav"
+            ]
+        }
+    },
+    {
+        "id": 7712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008340.wav"
+            ]
+        }
+    },
+    {
+        "id": 7713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002470.wav"
+            ]
+        }
+    },
+    {
+        "id": 7715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004360.wav"
+            ]
+        }
+    },
+    {
+        "id": 7716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007846.wav"
+            ]
+        }
+    },
+    {
+        "id": 7718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003283.wav"
+            ]
+        }
+    },
+    {
+        "id": 7725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000961.wav"
+            ]
+        }
+    },
+    {
+        "id": 7726,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001305.wav"
+            ]
+        }
+    },
+    {
+        "id": 7731,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003957.wav"
+            ]
+        }
+    },
+    {
+        "id": 7738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006191.wav"
+            ]
+        }
+    },
+    {
+        "id": 7740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001222.wav"
+            ]
+        }
+    },
+    {
+        "id": 7745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006719.wav"
+            ]
+        }
+    },
+    {
+        "id": 7750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007176.wav"
+            ]
+        }
+    },
+    {
+        "id": 7763,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006087.wav"
+            ]
+        }
+    },
+    {
+        "id": 7765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006003.wav"
+            ]
+        }
+    },
+    {
+        "id": 7767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005376.wav"
+            ]
+        }
+    },
+    {
+        "id": 7772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006620.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006620.wav"
+            ]
+        }
+    },
+    {
+        "id": 7774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000717.wav"
+            ]
+        }
+    },
+    {
+        "id": 7778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001233.wav"
+            ]
+        }
+    },
+    {
+        "id": 7792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005075.wav"
+            ]
+        }
+    },
+    {
+        "id": 7798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003520.wav"
+            ]
+        }
+    },
+    {
+        "id": 7803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004489.wav"
+            ]
+        }
+    },
+    {
+        "id": 7806,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008522.wav"
+            ]
+        }
+    },
+    {
+        "id": 7810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004460.wav"
+            ]
+        }
+    },
+    {
+        "id": 7821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007908.wav"
+            ]
+        }
+    },
+    {
+        "id": 7825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000294.wav"
+            ]
+        }
+    },
+    {
+        "id": 7826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006329.wav"
+            ]
+        }
+    },
+    {
+        "id": 7829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008173.wav"
+            ]
+        }
+    },
+    {
+        "id": 7835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003097.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003097.wav"
+            ]
+        }
+    },
+    {
+        "id": 7838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008551.wav"
+            ]
+        }
+    },
+    {
+        "id": 7839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000148.wav"
+            ]
+        }
+    },
+    {
+        "id": 7843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003441.wav"
+            ]
+        }
+    },
+    {
+        "id": 7849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006465.wav"
+            ]
+        }
+    },
+    {
+        "id": 7855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005547.wav"
+            ]
+        }
+    },
+    {
+        "id": 7859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007856.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007856.wav"
+            ]
+        }
+    },
+    {
+        "id": 7866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007825.wav"
+            ]
+        }
+    },
+    {
+        "id": 7872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006804.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006804.wav"
+            ]
+        }
+    },
+    {
+        "id": 7875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005064.wav"
+            ]
+        }
+    },
+    {
+        "id": 7890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008276.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008276.wav"
+            ]
+        }
+    },
+    {
+        "id": 7893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005386.wav"
+            ]
+        }
+    },
+    {
+        "id": 7895,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003606.wav"
+            ]
+        }
+    },
+    {
+        "id": 7897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003938.wav"
+            ]
+        }
+    },
+    {
+        "id": 7898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000977.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000977.wav"
+            ]
+        }
+    },
+    {
+        "id": 7905,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000539.wav"
+            ]
+        }
+    },
+    {
+        "id": 7925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005648.wav"
+            ]
+        }
+    },
+    {
+        "id": 7927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002758.wav"
+            ]
+        }
+    },
+    {
+        "id": 7935,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004563.wav"
+            ]
+        }
+    },
+    {
+        "id": 7936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006836.wav"
+            ]
+        }
+    },
+    {
+        "id": 7945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004616.wav"
+            ]
+        }
+    },
+    {
+        "id": 7949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002397.wav"
+            ]
+        }
+    },
+    {
+        "id": 7951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001571.wav"
+            ]
+        }
+    },
+    {
+        "id": 7952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005541.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005541.wav"
+            ]
+        }
+    },
+    {
+        "id": 7953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000968.wav"
+            ]
+        }
+    },
+    {
+        "id": 7958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001950.wav"
+            ]
+        }
+    },
+    {
+        "id": 7960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001446.wav"
+            ]
+        }
+    },
+    {
+        "id": 7964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005678.wav"
+            ]
+        }
+    },
+    {
+        "id": 7972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003989.wav"
+            ]
+        }
+    },
+    {
+        "id": 7983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002025.wav"
+            ]
+        }
+    },
+    {
+        "id": 7999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006861.wav"
+            ]
+        }
+    },
+    {
+        "id": 8001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000903.wav"
+            ]
+        }
+    },
+    {
+        "id": 8003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006310.wav"
+            ]
+        }
+    },
+    {
+        "id": 8007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000211.wav"
+            ]
+        }
+    },
+    {
+        "id": 8008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001697.wav"
+            ]
+        }
+    },
+    {
+        "id": 8016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001360.wav"
+            ]
+        }
+    },
+    {
+        "id": 8019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001386.wav"
+            ]
+        }
+    },
+    {
+        "id": 8020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007986.wav"
+            ]
+        }
+    },
+    {
+        "id": 8023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000425.wav"
+            ]
+        }
+    },
+    {
+        "id": 8024,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000556.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000556.wav"
+            ]
+        }
+    },
+    {
+        "id": 8026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004749.wav"
+            ]
+        }
+    },
+    {
+        "id": 8033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008954.wav"
+            ]
+        }
+    },
+    {
+        "id": 8046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006772.wav"
+            ]
+        }
+    },
+    {
+        "id": 8053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003087.wav"
+            ]
+        }
+    },
+    {
+        "id": 8059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003966.wav"
+            ]
+        }
+    },
+    {
+        "id": 8062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006911.wav"
+            ]
+        }
+    },
+    {
+        "id": 8063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004186.wav"
+            ]
+        }
+    },
+    {
+        "id": 8066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001640.wav"
+            ]
+        }
+    },
+    {
+        "id": 8077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004167.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004167.wav"
+            ]
+        }
+    },
+    {
+        "id": 8087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002073.wav"
+            ]
+        }
+    },
+    {
+        "id": 8088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left louder than the clarinet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007230.wav"
+            ]
+        }
+    },
+    {
+        "id": 8090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007224.wav"
+            ]
+        }
+    },
+    {
+        "id": 8092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004621.wav"
+            ]
+        }
+    },
+    {
+        "id": 8095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008236.wav"
+            ]
+        }
+    },
+    {
+        "id": 8096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004705.wav"
+            ]
+        }
+    },
+    {
+        "id": 8100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007798.wav"
+            ]
+        }
+    },
+    {
+        "id": 8102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003021.wav"
+            ]
+        }
+    },
+    {
+        "id": 8104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005420.wav"
+            ]
+        }
+    },
+    {
+        "id": 8105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004953.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004953.wav"
+            ]
+        }
+    },
+    {
+        "id": 8110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005598.wav"
+            ]
+        }
+    },
+    {
+        "id": 8115,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006860.wav"
+            ]
+        }
+    },
+    {
+        "id": 8118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006184.wav"
+            ]
+        }
+    },
+    {
+        "id": 8120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004990.wav"
+            ]
+        }
+    },
+    {
+        "id": 8132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000138.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000138.wav"
+            ]
+        }
+    },
+    {
+        "id": 8137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003037.wav"
+            ]
+        }
+    },
+    {
+        "id": 8139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the congas on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006854.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006854.wav"
+            ]
+        }
+    },
+    {
+        "id": 8153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007527.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007527.wav"
+            ]
+        }
+    },
+    {
+        "id": 8154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005910.wav"
+            ]
+        }
+    },
+    {
+        "id": 8156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001227.wav"
+            ]
+        }
+    },
+    {
+        "id": 8163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003973.wav"
+            ]
+        }
+    },
+    {
+        "id": 8176,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001746.wav"
+            ]
+        }
+    },
+    {
+        "id": 8180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002199.wav"
+            ]
+        }
+    },
+    {
+        "id": 8184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004357.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004357.wav"
+            ]
+        }
+    },
+    {
+        "id": 8189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003567.wav"
+            ]
+        }
+    },
+    {
+        "id": 8192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004750.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004750.wav"
+            ]
+        }
+    },
+    {
+        "id": 8197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003393.wav"
+            ]
+        }
+    },
+    {
+        "id": 8198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003017.wav"
+            ]
+        }
+    },
+    {
+        "id": 8200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005866.wav"
+            ]
+        }
+    },
+    {
+        "id": 8205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000033.wav"
+            ]
+        }
+    },
+    {
+        "id": 8213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001275.wav"
+            ]
+        }
+    },
+    {
+        "id": 8218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005916.wav"
+            ]
+        }
+    },
+    {
+        "id": 8220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002095.wav"
+            ]
+        }
+    },
+    {
+        "id": 8221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001737.wav"
+            ]
+        }
+    },
+    {
+        "id": 8224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000856.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000856.wav"
+            ]
+        }
+    },
+    {
+        "id": 8234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004092.wav"
+            ]
+        }
+    },
+    {
+        "id": 8237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002015.wav"
+            ]
+        }
+    },
+    {
+        "id": 8238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004994.wav"
+            ]
+        }
+    },
+    {
+        "id": 8242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008474.wav"
+            ]
+        }
+    },
+    {
+        "id": 8245,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006407.wav"
+            ]
+        }
+    },
+    {
+        "id": 8251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006123.wav"
+            ]
+        }
+    },
+    {
+        "id": 8254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007883.wav"
+            ]
+        }
+    },
+    {
+        "id": 8260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000628.wav"
+            ]
+        }
+    },
+    {
+        "id": 8270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003185.wav"
+            ]
+        }
+    },
+    {
+        "id": 8279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006132.wav"
+            ]
+        }
+    },
+    {
+        "id": 8283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007365.wav"
+            ]
+        }
+    },
+    {
+        "id": 8284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the right louder than the electric_bass on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004981.wav"
+            ]
+        }
+    },
+    {
+        "id": 8294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008386.wav"
+            ]
+        }
+    },
+    {
+        "id": 8304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000618.wav"
+            ]
+        }
+    },
+    {
+        "id": 8307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004992.wav"
+            ]
+        }
+    },
+    {
+        "id": 8315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005703.wav"
+            ]
+        }
+    },
+    {
+        "id": 8331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006546.wav"
+            ]
+        }
+    },
+    {
+        "id": 8335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004934.wav"
+            ]
+        }
+    },
+    {
+        "id": 8345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004897.wav"
+            ]
+        }
+    },
+    {
+        "id": 8347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005845.wav"
+            ]
+        }
+    },
+    {
+        "id": 8356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000275.wav"
+            ]
+        }
+    },
+    {
+        "id": 8357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008190.wav"
+            ]
+        }
+    },
+    {
+        "id": 8359,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007874.wav"
+            ]
+        }
+    },
+    {
+        "id": 8360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002761.wav"
+            ]
+        }
+    },
+    {
+        "id": 8361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001125.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001125.wav"
+            ]
+        }
+    },
+    {
+        "id": 8364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001863.wav"
+            ]
+        }
+    },
+    {
+        "id": 8371,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007459.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007459.wav"
+            ]
+        }
+    },
+    {
+        "id": 8376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005790.wav"
+            ]
+        }
+    },
+    {
+        "id": 8378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003687.wav"
+            ]
+        }
+    },
+    {
+        "id": 8381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000502.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000502.wav"
+            ]
+        }
+    },
+    {
+        "id": 8384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006955.wav"
+            ]
+        }
+    },
+    {
+        "id": 8398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000928.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000928.wav"
+            ]
+        }
+    },
+    {
+        "id": 8400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002920.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002920.wav"
+            ]
+        }
+    },
+    {
+        "id": 8401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone on the right more rhythmic than the xylophone on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005711.wav"
+            ]
+        }
+    },
+    {
+        "id": 8402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the drum on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001885.wav"
+            ]
+        }
+    },
+    {
+        "id": 8419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003851.wav"
+            ]
+        }
+    },
+    {
+        "id": 8428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007712.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007712.wav"
+            ]
+        }
+    },
+    {
+        "id": 8433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008186.wav"
+            ]
+        }
+    },
+    {
+        "id": 8434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005591.wav"
+            ]
+        }
+    },
+    {
+        "id": 8435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006840.wav"
+            ]
+        }
+    },
+    {
+        "id": 8436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008891.wav"
+            ]
+        }
+    },
+    {
+        "id": 8438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003191.wav"
+            ]
+        }
+    },
+    {
+        "id": 8440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002941.wav"
+            ]
+        }
+    },
+    {
+        "id": 8441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bagpipe makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007130.wav"
+            ]
+        }
+    },
+    {
+        "id": 8448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005986.wav"
+            ]
+        }
+    },
+    {
+        "id": 8455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007375.wav"
+            ]
+        }
+    },
+    {
+        "id": 8462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000207.wav"
+            ]
+        }
+    },
+    {
+        "id": 8466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002883.wav"
+            ]
+        }
+    },
+    {
+        "id": 8467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000941.wav"
+            ]
+        }
+    },
+    {
+        "id": 8470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000637.wav"
+            ]
+        }
+    },
+    {
+        "id": 8478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001931.wav"
+            ]
+        }
+    },
+    {
+        "id": 8487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003491.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003491.wav"
+            ]
+        }
+    },
+    {
+        "id": 8490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002524.wav"
+            ]
+        }
+    },
+    {
+        "id": 8492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005281.wav"
+            ]
+        }
+    },
+    {
+        "id": 8498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003667.wav"
+            ]
+        }
+    },
+    {
+        "id": 8500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008172.wav"
+            ]
+        }
+    },
+    {
+        "id": 8502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002493.wav"
+            ]
+        }
+    },
+    {
+        "id": 8505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002725.wav"
+            ]
+        }
+    },
+    {
+        "id": 8523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006250.wav"
+            ]
+        }
+    },
+    {
+        "id": 8526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000943.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000943.wav"
+            ]
+        }
+    },
+    {
+        "id": 8527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006678.wav"
+            ]
+        }
+    },
+    {
+        "id": 8531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003106.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003106.wav"
+            ]
+        }
+    },
+    {
+        "id": 8534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006194.wav"
+            ]
+        }
+    },
+    {
+        "id": 8535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008680.wav"
+            ]
+        }
+    },
+    {
+        "id": 8547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000290.wav"
+            ]
+        }
+    },
+    {
+        "id": 8563,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005066.wav"
+            ]
+        }
+    },
+    {
+        "id": 8571,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002517.wav"
+            ]
+        }
+    },
+    {
+        "id": 8578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the trumpet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002443.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002443.wav"
+            ]
+        }
+    },
+    {
+        "id": 8586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000559.wav"
+            ]
+        }
+    },
+    {
+        "id": 8587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003634.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003634.wav"
+            ]
+        }
+    },
+    {
+        "id": 8591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the fourth instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007836.wav"
+            ]
+        }
+    },
+    {
+        "id": 8604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001973.wav"
+            ]
+        }
+    },
+    {
+        "id": 8607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007402.wav"
+            ]
+        }
+    },
+    {
+        "id": 8618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003847.wav"
+            ]
+        }
+    },
+    {
+        "id": 8623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004784.wav"
+            ]
+        }
+    },
+    {
+        "id": 8624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001524.wav"
+            ]
+        }
+    },
+    {
+        "id": 8625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008020.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008020.wav"
+            ]
+        }
+    },
+    {
+        "id": 8628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003230.wav"
+            ]
+        }
+    },
+    {
+        "id": 8630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000662.wav"
+            ]
+        }
+    },
+    {
+        "id": 8637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007457.wav"
+            ]
+        }
+    },
+    {
+        "id": 8642,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003964.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003964.wav"
+            ]
+        }
+    },
+    {
+        "id": 8648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001387.wav"
+            ]
+        }
+    },
+    {
+        "id": 8649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004711.wav"
+            ]
+        }
+    },
+    {
+        "id": 8653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007567.wav"
+            ]
+        }
+    },
+    {
+        "id": 8656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003881.wav"
+            ]
+        }
+    },
+    {
+        "id": 8669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003014.wav"
+            ]
+        }
+    },
+    {
+        "id": 8671,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004860.wav"
+            ]
+        }
+    },
+    {
+        "id": 8673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007238.wav"
+            ]
+        }
+    },
+    {
+        "id": 8675,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004388.wav"
+            ]
+        }
+    },
+    {
+        "id": 8676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001623.wav"
+            ]
+        }
+    },
+    {
+        "id": 8688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008112.wav"
+            ]
+        }
+    },
+    {
+        "id": 8698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000309.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000309.wav"
+            ]
+        }
+    },
+    {
+        "id": 8699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001648.wav"
+            ]
+        }
+    },
+    {
+        "id": 8709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000705.wav"
+            ]
+        }
+    },
+    {
+        "id": 8711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000936.wav"
+            ]
+        }
+    },
+    {
+        "id": 8712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000787.wav"
+            ]
+        }
+    },
+    {
+        "id": 8713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006757.wav"
+            ]
+        }
+    },
+    {
+        "id": 8714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000378.wav"
+            ]
+        }
+    },
+    {
+        "id": 8715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007762.wav"
+            ]
+        }
+    },
+    {
+        "id": 8716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002658.wav"
+            ]
+        }
+    },
+    {
+        "id": 8718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005762.wav"
+            ]
+        }
+    },
+    {
+        "id": 8726,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003733.wav"
+            ]
+        }
+    },
+    {
+        "id": 8730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003733.wav"
+            ]
+        }
+    },
+    {
+        "id": 8733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007062.wav"
+            ]
+        }
+    },
+    {
+        "id": 8734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008685.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008685.wav"
+            ]
+        }
+    },
+    {
+        "id": 8739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001172.wav"
+            ]
+        }
+    },
+    {
+        "id": 8741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005254.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005254.wav"
+            ]
+        }
+    },
+    {
+        "id": 8748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003599.wav"
+            ]
+        }
+    },
+    {
+        "id": 8749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001682.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001682.wav"
+            ]
+        }
+    },
+    {
+        "id": 8756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002843.wav"
+            ]
+        }
+    },
+    {
+        "id": 8762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004506.wav"
+            ]
+        }
+    },
+    {
+        "id": 8765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000462.wav"
+            ]
+        }
+    },
+    {
+        "id": 8772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006948.wav"
+            ]
+        }
+    },
+    {
+        "id": 8785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002528.wav"
+            ]
+        }
+    },
+    {
+        "id": 8792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001596.wav"
+            ]
+        }
+    },
+    {
+        "id": 8793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007990.wav"
+            ]
+        }
+    },
+    {
+        "id": 8794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006693.wav"
+            ]
+        }
+    },
+    {
+        "id": 8799,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003534.wav"
+            ]
+        }
+    },
+    {
+        "id": 8802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007480.wav"
+            ]
+        }
+    },
+    {
+        "id": 8811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005188.wav"
+            ]
+        }
+    },
+    {
+        "id": 8812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006704.wav"
+            ]
+        }
+    },
+    {
+        "id": 8813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001055.wav"
+            ]
+        }
+    },
+    {
+        "id": 8815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001361.wav"
+            ]
+        }
+    },
+    {
+        "id": 8820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006928.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006928.wav"
+            ]
+        }
+    },
+    {
+        "id": 8823,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007615.wav"
+            ]
+        }
+    },
+    {
+        "id": 8829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008562.wav"
+            ]
+        }
+    },
+    {
+        "id": 8833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006333.wav"
+            ]
+        }
+    },
+    {
+        "id": 8836,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001491.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001491.wav"
+            ]
+        }
+    },
+    {
+        "id": 8842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005294.wav"
+            ]
+        }
+    },
+    {
+        "id": 8856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008241.wav"
+            ]
+        }
+    },
+    {
+        "id": 8877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002901.wav"
+            ]
+        }
+    },
+    {
+        "id": 8880,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001861.wav"
+            ]
+        }
+    },
+    {
+        "id": 8883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007260.wav"
+            ]
+        }
+    },
+    {
+        "id": 8889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000631.wav"
+            ]
+        }
+    },
+    {
+        "id": 8900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000222.wav"
+            ]
+        }
+    },
+    {
+        "id": 8906,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008370.wav"
+            ]
+        }
+    },
+    {
+        "id": 8911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003190.wav"
+            ]
+        }
+    },
+    {
+        "id": 8920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004397.wav"
+            ]
+        }
+    },
+    {
+        "id": 8926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003274.wav"
+            ]
+        }
+    },
+    {
+        "id": 8928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003424.wav"
+            ]
+        }
+    },
+    {
+        "id": 8929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005487.wav"
+            ]
+        }
+    },
+    {
+        "id": 8932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003865.wav"
+            ]
+        }
+    },
+    {
+        "id": 8940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005582.wav"
+            ]
+        }
+    },
+    {
+        "id": 8943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004737.wav"
+            ]
+        }
+    },
+    {
+        "id": 8957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004085.wav"
+            ]
+        }
+    },
+    {
+        "id": 8964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004738.wav"
+            ]
+        }
+    },
+    {
+        "id": 8972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005017.wav"
+            ]
+        }
+    },
+    {
+        "id": 8978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001085.wav"
+            ]
+        }
+    },
+    {
+        "id": 8979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008714.wav"
+            ]
+        }
+    },
+    {
+        "id": 8985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the fourth instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005163.wav"
+            ]
+        }
+    },
+    {
+        "id": 8986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002857.wav"
+            ]
+        }
+    },
+    {
+        "id": 8989,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004782.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004782.wav"
+            ]
+        }
+    },
+    {
+        "id": 8997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005774.wav"
+            ]
+        }
+    },
+    {
+        "id": 8998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008294.wav"
+            ]
+        }
+    },
+    {
+        "id": 9001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008712.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008712.wav"
+            ]
+        }
+    },
+    {
+        "id": 9009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone on the middle louder than the xylophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000346.wav"
+            ]
+        }
+    },
+    {
+        "id": 9011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006760.wav"
+            ]
+        }
+    },
+    {
+        "id": 9013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007492.wav"
+            ]
+        }
+    },
+    {
+        "id": 9018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005618.wav"
+            ]
+        }
+    },
+    {
+        "id": 9022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003877.wav"
+            ]
+        }
+    },
+    {
+        "id": 9023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005901.wav"
+            ]
+        }
+    },
+    {
+        "id": 9028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007938.wav"
+            ]
+        }
+    },
+    {
+        "id": 9029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004089.wav"
+            ]
+        }
+    },
+    {
+        "id": 9031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002438.wav"
+            ]
+        }
+    },
+    {
+        "id": 9032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the left louder than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003714.wav"
+            ]
+        }
+    },
+    {
+        "id": 9034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001588.wav"
+            ]
+        }
+    },
+    {
+        "id": 9037,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008326.wav"
+            ]
+        }
+    },
+    {
+        "id": 9048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003119.wav"
+            ]
+        }
+    },
+    {
+        "id": 9065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001711.wav"
+            ]
+        }
+    },
+    {
+        "id": 9075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004698.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004698.wav"
+            ]
+        }
+    },
+    {
+        "id": 9081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000818.wav"
+            ]
+        }
+    },
+    {
+        "id": 9089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004839.wav"
+            ]
+        }
+    },
+    {
+        "id": 9095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008621.wav"
+            ]
+        }
+    },
+    {
+        "id": 9098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001015.wav"
+            ]
+        }
+    },
+    {
+        "id": 9105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005392.wav"
+            ]
+        }
+    },
+    {
+        "id": 9107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003389.wav"
+            ]
+        }
+    },
+    {
+        "id": 9120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000759.wav"
+            ]
+        }
+    },
+    {
+        "id": 9123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001891.wav"
+            ]
+        }
+    },
+    {
+        "id": 9126,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006676.wav"
+            ]
+        }
+    },
+    {
+        "id": 9128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004361.wav"
+            ]
+        }
+    },
+    {
+        "id": 9135,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001139.wav"
+            ]
+        }
+    },
+    {
+        "id": 9138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002280.wav"
+            ]
+        }
+    },
+    {
+        "id": 9143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the fifth instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006695.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006695.wav"
+            ]
+        }
+    },
+    {
+        "id": 9147,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007837.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007837.wav"
+            ]
+        }
+    },
+    {
+        "id": 9152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the clarinet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003892.wav"
+            ]
+        }
+    },
+    {
+        "id": 9155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005072.wav"
+            ]
+        }
+    },
+    {
+        "id": 9162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003519.wav"
+            ]
+        }
+    },
+    {
+        "id": 9165,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002684.wav"
+            ]
+        }
+    },
+    {
+        "id": 9167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006275.wav"
+            ]
+        }
+    },
+    {
+        "id": 9176,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005309.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005309.wav"
+            ]
+        }
+    },
+    {
+        "id": 9186,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003169.wav"
+            ]
+        }
+    },
+    {
+        "id": 9189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004427.wav"
+            ]
+        }
+    },
+    {
+        "id": 9192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004744.wav"
+            ]
+        }
+    },
+    {
+        "id": 9202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006839.wav"
+            ]
+        }
+    },
+    {
+        "id": 9204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008800.wav"
+            ]
+        }
+    },
+    {
+        "id": 9208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002739.wav"
+            ]
+        }
+    },
+    {
+        "id": 9217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right more rhythmic than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006572.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006572.wav"
+            ]
+        }
+    },
+    {
+        "id": 9218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005940.wav"
+            ]
+        }
+    },
+    {
+        "id": 9219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000783.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000783.wav"
+            ]
+        }
+    },
+    {
+        "id": 9223,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006804.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006804.wav"
+            ]
+        }
+    },
+    {
+        "id": 9237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006816.wav"
+            ]
+        }
+    },
+    {
+        "id": 9238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008613.wav"
+            ]
+        }
+    },
+    {
+        "id": 9247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002041.wav"
+            ]
+        }
+    },
+    {
+        "id": 9250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003710.wav"
+            ]
+        }
+    },
+    {
+        "id": 9254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006656.wav"
+            ]
+        }
+    },
+    {
+        "id": 9265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004768.wav"
+            ]
+        }
+    },
+    {
+        "id": 9273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000675.wav"
+            ]
+        }
+    },
+    {
+        "id": 9275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008310.wav"
+            ]
+        }
+    },
+    {
+        "id": 9283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004678.wav"
+            ]
+        }
+    },
+    {
+        "id": 9284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000425.wav"
+            ]
+        }
+    },
+    {
+        "id": 9287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008356.wav"
+            ]
+        }
+    },
+    {
+        "id": 9293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004496.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004496.wav"
+            ]
+        }
+    },
+    {
+        "id": 9296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005962.wav"
+            ]
+        }
+    },
+    {
+        "id": 9304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004652.wav"
+            ]
+        }
+    },
+    {
+        "id": 9313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000208.wav"
+            ]
+        }
+    },
+    {
+        "id": 9314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003111.wav"
+            ]
+        }
+    },
+    {
+        "id": 9315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008856.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008856.wav"
+            ]
+        }
+    },
+    {
+        "id": 9316,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005905.wav"
+            ]
+        }
+    },
+    {
+        "id": 9317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005562.wav"
+            ]
+        }
+    },
+    {
+        "id": 9323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007471.wav"
+            ]
+        }
+    },
+    {
+        "id": 9325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001874.wav"
+            ]
+        }
+    },
+    {
+        "id": 9329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004740.wav"
+            ]
+        }
+    },
+    {
+        "id": 9333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001092.wav"
+            ]
+        }
+    },
+    {
+        "id": 9346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004816.wav"
+            ]
+        }
+    },
+    {
+        "id": 9350,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006994.wav"
+            ]
+        }
+    },
+    {
+        "id": 9351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002967.wav"
+            ]
+        }
+    },
+    {
+        "id": 9361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003188.wav"
+            ]
+        }
+    },
+    {
+        "id": 9362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000281.wav"
+            ]
+        }
+    },
+    {
+        "id": 9363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003066.wav"
+            ]
+        }
+    },
+    {
+        "id": 9364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005456.wav"
+            ]
+        }
+    },
+    {
+        "id": 9365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000895.wav"
+            ]
+        }
+    },
+    {
+        "id": 9368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006708.wav"
+            ]
+        }
+    },
+    {
+        "id": 9370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007792.wav"
+            ]
+        }
+    },
+    {
+        "id": 9380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007501.wav"
+            ]
+        }
+    },
+    {
+        "id": 9382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002971.wav"
+            ]
+        }
+    },
+    {
+        "id": 9386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the right more rhythmic than the banjo on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006584.wav"
+            ]
+        }
+    },
+    {
+        "id": 9391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003029.wav"
+            ]
+        }
+    },
+    {
+        "id": 9397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007989.wav"
+            ]
+        }
+    },
+    {
+        "id": 9399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005435.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005435.wav"
+            ]
+        }
+    },
+    {
+        "id": 9402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006743.wav"
+            ]
+        }
+    },
+    {
+        "id": 9409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich banjo makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007291.wav"
+            ]
+        }
+    },
+    {
+        "id": 9412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005992.wav"
+            ]
+        }
+    },
+    {
+        "id": 9420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002947.wav"
+            ]
+        }
+    },
+    {
+        "id": 9422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008366.wav"
+            ]
+        }
+    },
+    {
+        "id": 9423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001573.wav"
+            ]
+        }
+    },
+    {
+        "id": 9434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004699.wav"
+            ]
+        }
+    },
+    {
+        "id": 9441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007740.wav"
+            ]
+        }
+    },
+    {
+        "id": 9445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich xylophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001930.wav"
+            ]
+        }
+    },
+    {
+        "id": 9446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003499.wav"
+            ]
+        }
+    },
+    {
+        "id": 9459,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004192.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004192.wav"
+            ]
+        }
+    },
+    {
+        "id": 9464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000054.wav"
+            ]
+        }
+    },
+    {
+        "id": 9465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004079.wav"
+            ]
+        }
+    },
+    {
+        "id": 9468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000432.wav"
+            ]
+        }
+    },
+    {
+        "id": 9470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003000.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003000.wav"
+            ]
+        }
+    },
+    {
+        "id": 9472,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002419.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002419.wav"
+            ]
+        }
+    },
+    {
+        "id": 9480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005300.wav"
+            ]
+        }
+    },
+    {
+        "id": 9483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001476.wav"
+            ]
+        }
+    },
+    {
+        "id": 9484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004806.wav"
+            ]
+        }
+    },
+    {
+        "id": 9487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008251.wav"
+            ]
+        }
+    },
+    {
+        "id": 9491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004600.wav"
+            ]
+        }
+    },
+    {
+        "id": 9494,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008506.wav"
+            ]
+        }
+    },
+    {
+        "id": 9496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003308.wav"
+            ]
+        }
+    },
+    {
+        "id": 9498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005268.wav"
+            ]
+        }
+    },
+    {
+        "id": 9505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005502.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005502.wav"
+            ]
+        }
+    },
+    {
+        "id": 9506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005742.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005742.wav"
+            ]
+        }
+    },
+    {
+        "id": 9508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008873.wav"
+            ]
+        }
+    },
+    {
+        "id": 9509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002413.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002413.wav"
+            ]
+        }
+    },
+    {
+        "id": 9513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007517.wav"
+            ]
+        }
+    },
+    {
+        "id": 9514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001725.wav"
+            ]
+        }
+    },
+    {
+        "id": 9519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000004.wav"
+            ]
+        }
+    },
+    {
+        "id": 9520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002143.wav"
+            ]
+        }
+    },
+    {
+        "id": 9524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005253.wav"
+            ]
+        }
+    },
+    {
+        "id": 9525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004006.wav"
+            ]
+        }
+    },
+    {
+        "id": 9532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002186.wav"
+            ]
+        }
+    },
+    {
+        "id": 9536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004572.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004572.wav"
+            ]
+        }
+    },
+    {
+        "id": 9551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007652.wav"
+            ]
+        }
+    },
+    {
+        "id": 9553,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003730.wav"
+            ]
+        }
+    },
+    {
+        "id": 9566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005500.wav"
+            ]
+        }
+    },
+    {
+        "id": 9569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005392.wav"
+            ]
+        }
+    },
+    {
+        "id": 9576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004890.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004890.wav"
+            ]
+        }
+    },
+    {
+        "id": 9579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003528.wav"
+            ]
+        }
+    },
+    {
+        "id": 9582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003469.wav"
+            ]
+        }
+    },
+    {
+        "id": 9583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001758.wav"
+            ]
+        }
+    },
+    {
+        "id": 9593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007280.wav"
+            ]
+        }
+    },
+    {
+        "id": 9596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006270.wav"
+            ]
+        }
+    },
+    {
+        "id": 9598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008090.wav"
+            ]
+        }
+    },
+    {
+        "id": 9603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003418.wav"
+            ]
+        }
+    },
+    {
+        "id": 9608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008514.wav"
+            ]
+        }
+    },
+    {
+        "id": 9618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006727.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006727.wav"
+            ]
+        }
+    },
+    {
+        "id": 9621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002950.wav"
+            ]
+        }
+    },
+    {
+        "id": 9624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007857.wav"
+            ]
+        }
+    },
+    {
+        "id": 9626,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005866.wav"
+            ]
+        }
+    },
+    {
+        "id": 9627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005898.wav"
+            ]
+        }
+    },
+    {
+        "id": 9628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007579.wav"
+            ]
+        }
+    },
+    {
+        "id": 9629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005082.wav"
+            ]
+        }
+    },
+    {
+        "id": 9643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005743.wav"
+            ]
+        }
+    },
+    {
+        "id": 9644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001555.wav"
+            ]
+        }
+    },
+    {
+        "id": 9648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003714.wav"
+            ]
+        }
+    },
+    {
+        "id": 9649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002601.wav"
+            ]
+        }
+    },
+    {
+        "id": 9664,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004317.wav"
+            ]
+        }
+    },
+    {
+        "id": 9666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004973.wav"
+            ]
+        }
+    },
+    {
+        "id": 9670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005751.wav"
+            ]
+        }
+    },
+    {
+        "id": 9680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008058.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008058.wav"
+            ]
+        }
+    },
+    {
+        "id": 9682,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001240.wav"
+            ]
+        }
+    },
+    {
+        "id": 9686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001516.wav"
+            ]
+        }
+    },
+    {
+        "id": 9697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002071.wav"
+            ]
+        }
+    },
+    {
+        "id": 9698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001070.wav"
+            ]
+        }
+    },
+    {
+        "id": 9699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008766.wav"
+            ]
+        }
+    },
+    {
+        "id": 9701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008730.wav"
+            ]
+        }
+    },
+    {
+        "id": 9709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002506.wav"
+            ]
+        }
+    },
+    {
+        "id": 9713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005652.wav"
+            ]
+        }
+    },
+    {
+        "id": 9715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007360.wav"
+            ]
+        }
+    },
+    {
+        "id": 9717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005822.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005822.wav"
+            ]
+        }
+    },
+    {
+        "id": 9722,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004815.wav"
+            ]
+        }
+    },
+    {
+        "id": 9728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004350.wav"
+            ]
+        }
+    },
+    {
+        "id": 9741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000797.wav"
+            ]
+        }
+    },
+    {
+        "id": 9742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005962.wav"
+            ]
+        }
+    },
+    {
+        "id": 9755,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004679.wav"
+            ]
+        }
+    },
+    {
+        "id": 9762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008799.wav"
+            ]
+        }
+    },
+    {
+        "id": 9776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001532.wav"
+            ]
+        }
+    },
+    {
+        "id": 9779,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005132.wav"
+            ]
+        }
+    },
+    {
+        "id": 9786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002575.wav"
+            ]
+        }
+    },
+    {
+        "id": 9790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004424.wav"
+            ]
+        }
+    },
+    {
+        "id": 9792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000385.wav"
+            ]
+        }
+    },
+    {
+        "id": 9807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001021.wav"
+            ]
+        }
+    },
+    {
+        "id": 9810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007660.wav"
+            ]
+        }
+    },
+    {
+        "id": 9811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008373.wav"
+            ]
+        }
+    },
+    {
+        "id": 9821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002389.wav"
+            ]
+        }
+    },
+    {
+        "id": 9822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001159.wav"
+            ]
+        }
+    },
+    {
+        "id": 9826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001778.wav"
+            ]
+        }
+    },
+    {
+        "id": 9827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005158.wav"
+            ]
+        }
+    },
+    {
+        "id": 9828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006315.wav"
+            ]
+        }
+    },
+    {
+        "id": 9833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000898.wav"
+            ]
+        }
+    },
+    {
+        "id": 9834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007071.wav"
+            ]
+        }
+    },
+    {
+        "id": 9838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001907.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001907.wav"
+            ]
+        }
+    },
+    {
+        "id": 9839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007773.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007773.wav"
+            ]
+        }
+    },
+    {
+        "id": 9847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002550.wav"
+            ]
+        }
+    },
+    {
+        "id": 9854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008635.wav"
+            ]
+        }
+    },
+    {
+        "id": 9856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006349.wav"
+            ]
+        }
+    },
+    {
+        "id": 9859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008474.wav"
+            ]
+        }
+    },
+    {
+        "id": 9862,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006331.wav"
+            ]
+        }
+    },
+    {
+        "id": 9867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004970.wav"
+            ]
+        }
+    },
+    {
+        "id": 9868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001629.wav"
+            ]
+        }
+    },
+    {
+        "id": 9871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007811.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007811.wav"
+            ]
+        }
+    },
+    {
+        "id": 9879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002671.wav"
+            ]
+        }
+    },
+    {
+        "id": 9882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006670.wav"
+            ]
+        }
+    },
+    {
+        "id": 9892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002027.wav"
+            ]
+        }
+    },
+    {
+        "id": 9896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007409.wav"
+            ]
+        }
+    },
+    {
+        "id": 9914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004529.wav"
+            ]
+        }
+    },
+    {
+        "id": 9919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004133.wav"
+            ]
+        }
+    },
+    {
+        "id": 9926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003772.wav"
+            ]
+        }
+    },
+    {
+        "id": 9930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008718.wav"
+            ]
+        }
+    },
+    {
+        "id": 9932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007908.wav"
+            ]
+        }
+    },
+    {
+        "id": 9935,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007273.wav"
+            ]
+        }
+    },
+    {
+        "id": 9939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003599.wav"
+            ]
+        }
+    },
+    {
+        "id": 9944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005357.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005357.wav"
+            ]
+        }
+    },
+    {
+        "id": 9953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002313.wav"
+            ]
+        }
+    },
+    {
+        "id": 9961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006280.wav"
+            ]
+        }
+    },
+    {
+        "id": 9966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002298.wav"
+            ]
+        }
+    },
+    {
+        "id": 9975,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004695.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004695.wav"
+            ]
+        }
+    },
+    {
+        "id": 9990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002321.wav"
+            ]
+        }
+    },
+    {
+        "id": 10004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right more rhythmic than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003997.wav"
+            ]
+        }
+    },
+    {
+        "id": 10018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001758.wav"
+            ]
+        }
+    },
+    {
+        "id": 10023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right more rhythmic than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001101.wav"
+            ]
+        }
+    },
+    {
+        "id": 10027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003432.wav"
+            ]
+        }
+    },
+    {
+        "id": 10034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000285.wav"
+            ]
+        }
+    },
+    {
+        "id": 10043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005246.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005246.wav"
+            ]
+        }
+    },
+    {
+        "id": 10050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006997.wav"
+            ]
+        }
+    },
+    {
+        "id": 10057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005517.wav"
+            ]
+        }
+    },
+    {
+        "id": 10061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005008.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005008.wav"
+            ]
+        }
+    },
+    {
+        "id": 10083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006537.wav"
+            ]
+        }
+    },
+    {
+        "id": 10109,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007551.wav"
+            ]
+        }
+    },
+    {
+        "id": 10113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003784.wav"
+            ]
+        }
+    },
+    {
+        "id": 10123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001654.wav"
+            ]
+        }
+    },
+    {
+        "id": 10125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005546.wav"
+            ]
+        }
+    },
+    {
+        "id": 10126,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003436.wav"
+            ]
+        }
+    },
+    {
+        "id": 10128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008185.wav"
+            ]
+        }
+    },
+    {
+        "id": 10140,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002434.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002434.wav"
+            ]
+        }
+    },
+    {
+        "id": 10141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001248.wav"
+            ]
+        }
+    },
+    {
+        "id": 10143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008449.wav"
+            ]
+        }
+    },
+    {
+        "id": 10158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003776.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003776.wav"
+            ]
+        }
+    },
+    {
+        "id": 10167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000542.wav"
+            ]
+        }
+    },
+    {
+        "id": 10173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006253.wav"
+            ]
+        }
+    },
+    {
+        "id": 10174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007722.wav"
+            ]
+        }
+    },
+    {
+        "id": 10180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007656.wav"
+            ]
+        }
+    },
+    {
+        "id": 10181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005702.wav"
+            ]
+        }
+    },
+    {
+        "id": 10183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001744.wav"
+            ]
+        }
+    },
+    {
+        "id": 10198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003657.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003657.wav"
+            ]
+        }
+    },
+    {
+        "id": 10199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001333.wav"
+            ]
+        }
+    },
+    {
+        "id": 10207,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003436.wav"
+            ]
+        }
+    },
+    {
+        "id": 10211,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000126.wav"
+            ]
+        }
+    },
+    {
+        "id": 10217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005665.wav"
+            ]
+        }
+    },
+    {
+        "id": 10219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008261.wav"
+            ]
+        }
+    },
+    {
+        "id": 10236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000759.wav"
+            ]
+        }
+    },
+    {
+        "id": 10239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008301.wav"
+            ]
+        }
+    },
+    {
+        "id": 10248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007671.wav"
+            ]
+        }
+    },
+    {
+        "id": 10249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008220.wav"
+            ]
+        }
+    },
+    {
+        "id": 10265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007857.wav"
+            ]
+        }
+    },
+    {
+        "id": 10267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003544.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003544.wav"
+            ]
+        }
+    },
+    {
+        "id": 10270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001222.wav"
+            ]
+        }
+    },
+    {
+        "id": 10280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001184.wav"
+            ]
+        }
+    },
+    {
+        "id": 10286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006355.wav"
+            ]
+        }
+    },
+    {
+        "id": 10291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007955.wav"
+            ]
+        }
+    },
+    {
+        "id": 10297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001036.wav"
+            ]
+        }
+    },
+    {
+        "id": 10299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007700.wav"
+            ]
+        }
+    },
+    {
+        "id": 10309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002316.wav"
+            ]
+        }
+    },
+    {
+        "id": 10318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000571.wav"
+            ]
+        }
+    },
+    {
+        "id": 10322,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002376.wav"
+            ]
+        }
+    },
+    {
+        "id": 10347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007499.wav"
+            ]
+        }
+    },
+    {
+        "id": 10360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005236.wav"
+            ]
+        }
+    },
+    {
+        "id": 10363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005034.wav"
+            ]
+        }
+    },
+    {
+        "id": 10373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001999.wav"
+            ]
+        }
+    },
+    {
+        "id": 10375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002952.wav"
+            ]
+        }
+    },
+    {
+        "id": 10384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002155.wav"
+            ]
+        }
+    },
+    {
+        "id": 10385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007198.wav"
+            ]
+        }
+    },
+    {
+        "id": 10391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004940.wav"
+            ]
+        }
+    },
+    {
+        "id": 10404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002409.wav"
+            ]
+        }
+    },
+    {
+        "id": 10406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007528.wav"
+            ]
+        }
+    },
+    {
+        "id": 10408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003325.wav"
+            ]
+        }
+    },
+    {
+        "id": 10423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002648.wav"
+            ]
+        }
+    },
+    {
+        "id": 10434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007418.wav"
+            ]
+        }
+    },
+    {
+        "id": 10441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008397.wav"
+            ]
+        }
+    },
+    {
+        "id": 10446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000512.wav"
+            ]
+        }
+    },
+    {
+        "id": 10457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007958.wav"
+            ]
+        }
+    },
+    {
+        "id": 10464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006004.wav"
+            ]
+        }
+    },
+    {
+        "id": 10475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001728.wav"
+            ]
+        }
+    },
+    {
+        "id": 10488,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004062.wav"
+            ]
+        }
+    },
+    {
+        "id": 10494,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005256.wav"
+            ]
+        }
+    },
+    {
+        "id": 10498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007290.wav"
+            ]
+        }
+    },
+    {
+        "id": 10503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006408.wav"
+            ]
+        }
+    },
+    {
+        "id": 10508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004403.wav"
+            ]
+        }
+    },
+    {
+        "id": 10534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004552.wav"
+            ]
+        }
+    },
+    {
+        "id": 10538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left louder than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008820.wav"
+            ]
+        }
+    },
+    {
+        "id": 10544,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008944.wav"
+            ]
+        }
+    },
+    {
+        "id": 10546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the right more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008257.wav"
+            ]
+        }
+    },
+    {
+        "id": 10559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000592.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000592.wav"
+            ]
+        }
+    },
+    {
+        "id": 10560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000662.wav"
+            ]
+        }
+    },
+    {
+        "id": 10561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001175.wav"
+            ]
+        }
+    },
+    {
+        "id": 10562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002926.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002926.wav"
+            ]
+        }
+    },
+    {
+        "id": 10567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the right more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006796.wav"
+            ]
+        }
+    },
+    {
+        "id": 10581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005492.wav"
+            ]
+        }
+    },
+    {
+        "id": 10584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003999.wav"
+            ]
+        }
+    },
+    {
+        "id": 10587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005732.wav"
+            ]
+        }
+    },
+    {
+        "id": 10617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002959.wav"
+            ]
+        }
+    },
+    {
+        "id": 10620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002091.wav"
+            ]
+        }
+    },
+    {
+        "id": 10621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002228.wav"
+            ]
+        }
+    },
+    {
+        "id": 10624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000017.wav"
+            ]
+        }
+    },
+    {
+        "id": 10627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000352.wav"
+            ]
+        }
+    },
+    {
+        "id": 10629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006617.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006617.wav"
+            ]
+        }
+    },
+    {
+        "id": 10633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006132.wav"
+            ]
+        }
+    },
+    {
+        "id": 10638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005067.wav"
+            ]
+        }
+    },
+    {
+        "id": 10643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004942.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004942.wav"
+            ]
+        }
+    },
+    {
+        "id": 10644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003875.wav"
+            ]
+        }
+    },
+    {
+        "id": 10649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000115.wav"
+            ]
+        }
+    },
+    {
+        "id": 10653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007131.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007131.wav"
+            ]
+        }
+    },
+    {
+        "id": 10654,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007969.wav"
+            ]
+        }
+    },
+    {
+        "id": 10660,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002168.wav"
+            ]
+        }
+    },
+    {
+        "id": 10662,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008629.wav"
+            ]
+        }
+    },
+    {
+        "id": 10666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007752.wav"
+            ]
+        }
+    },
+    {
+        "id": 10667,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005145.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005145.wav"
+            ]
+        }
+    },
+    {
+        "id": 10677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004790.wav"
+            ]
+        }
+    },
+    {
+        "id": 10683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003032.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003032.wav"
+            ]
+        }
+    },
+    {
+        "id": 10690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007387.wav"
+            ]
+        }
+    },
+    {
+        "id": 10702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001335.wav"
+            ]
+        }
+    },
+    {
+        "id": 10712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005893.wav"
+            ]
+        }
+    },
+    {
+        "id": 10722,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005367.wav"
+            ]
+        }
+    },
+    {
+        "id": 10728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the middle more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004678.wav"
+            ]
+        }
+    },
+    {
+        "id": 10732,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003129.wav"
+            ]
+        }
+    },
+    {
+        "id": 10738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007615.wav"
+            ]
+        }
+    },
+    {
+        "id": 10740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001489.wav"
+            ]
+        }
+    },
+    {
+        "id": 10749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002684.wav"
+            ]
+        }
+    },
+    {
+        "id": 10759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the saxophone on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004933.wav"
+            ]
+        }
+    },
+    {
+        "id": 10760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005582.wav"
+            ]
+        }
+    },
+    {
+        "id": 10764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004876.wav"
+            ]
+        }
+    },
+    {
+        "id": 10769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left louder than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003030.wav"
+            ]
+        }
+    },
+    {
+        "id": 10774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002552.wav"
+            ]
+        }
+    },
+    {
+        "id": 10785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004206.wav"
+            ]
+        }
+    },
+    {
+        "id": 10787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001117.wav"
+            ]
+        }
+    },
+    {
+        "id": 10788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001164.wav"
+            ]
+        }
+    },
+    {
+        "id": 10789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001935.wav"
+            ]
+        }
+    },
+    {
+        "id": 10811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005562.wav"
+            ]
+        }
+    },
+    {
+        "id": 10816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003304.wav"
+            ]
+        }
+    },
+    {
+        "id": 10825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007537.wav"
+            ]
+        }
+    },
+    {
+        "id": 10834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000778.wav"
+            ]
+        }
+    },
+    {
+        "id": 10837,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006265.wav"
+            ]
+        }
+    },
+    {
+        "id": 10846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008069.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008069.wav"
+            ]
+        }
+    },
+    {
+        "id": 10855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006359.wav"
+            ]
+        }
+    },
+    {
+        "id": 10857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007547.wav"
+            ]
+        }
+    },
+    {
+        "id": 10861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002076.wav"
+            ]
+        }
+    },
+    {
+        "id": 10865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006287.wav"
+            ]
+        }
+    },
+    {
+        "id": 10873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002353.wav"
+            ]
+        }
+    },
+    {
+        "id": 10875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003275.wav"
+            ]
+        }
+    },
+    {
+        "id": 10878,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002839.wav"
+            ]
+        }
+    },
+    {
+        "id": 10882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000041.wav"
+            ]
+        }
+    },
+    {
+        "id": 10889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the right more rhythmic than the trumpet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005170.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005170.wav"
+            ]
+        }
+    },
+    {
+        "id": 10892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000890.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000890.wav"
+            ]
+        }
+    },
+    {
+        "id": 10897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002427.wav"
+            ]
+        }
+    },
+    {
+        "id": 10902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006066.wav"
+            ]
+        }
+    },
+    {
+        "id": 10912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006369.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006369.wav"
+            ]
+        }
+    },
+    {
+        "id": 10925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the banjo on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000767.wav"
+            ]
+        }
+    },
+    {
+        "id": 10928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004527.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004527.wav"
+            ]
+        }
+    },
+    {
+        "id": 10932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005442.wav"
+            ]
+        }
+    },
+    {
+        "id": 10944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001112.wav"
+            ]
+        }
+    },
+    {
+        "id": 10946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008231.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008231.wav"
+            ]
+        }
+    },
+    {
+        "id": 10947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008150.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008150.wav"
+            ]
+        }
+    },
+    {
+        "id": 10961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005531.wav"
+            ]
+        }
+    },
+    {
+        "id": 10972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004986.wav"
+            ]
+        }
+    },
+    {
+        "id": 10974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002997.wav"
+            ]
+        }
+    },
+    {
+        "id": 10976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007320.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007320.wav"
+            ]
+        }
+    },
+    {
+        "id": 10978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001193.wav"
+            ]
+        }
+    },
+    {
+        "id": 10981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006072.wav"
+            ]
+        }
+    },
+    {
+        "id": 10997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003058.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003058.wav"
+            ]
+        }
+    },
+    {
+        "id": 11011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004416.wav"
+            ]
+        }
+    },
+    {
+        "id": 11020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003060.wav"
+            ]
+        }
+    },
+    {
+        "id": 11025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006212.wav"
+            ]
+        }
+    },
+    {
+        "id": 11031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007025.wav"
+            ]
+        }
+    },
+    {
+        "id": 11034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007510.wav"
+            ]
+        }
+    },
+    {
+        "id": 11036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004631.wav"
+            ]
+        }
+    },
+    {
+        "id": 11040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002730.wav"
+            ]
+        }
+    },
+    {
+        "id": 11042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003025.wav"
+            ]
+        }
+    },
+    {
+        "id": 11056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003045.wav"
+            ]
+        }
+    },
+    {
+        "id": 11058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006075.wav"
+            ]
+        }
+    },
+    {
+        "id": 11066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000656.wav"
+            ]
+        }
+    },
+    {
+        "id": 11071,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004803.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004803.wav"
+            ]
+        }
+    },
+    {
+        "id": 11073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006735.wav"
+            ]
+        }
+    },
+    {
+        "id": 11077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003380.wav"
+            ]
+        }
+    },
+    {
+        "id": 11093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007078.wav"
+            ]
+        }
+    },
+    {
+        "id": 11111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the middle more rhythmic than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007269.wav"
+            ]
+        }
+    },
+    {
+        "id": 11112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003007.wav"
+            ]
+        }
+    },
+    {
+        "id": 11115,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004834.wav"
+            ]
+        }
+    },
+    {
+        "id": 11116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003000.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003000.wav"
+            ]
+        }
+    },
+    {
+        "id": 11117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008894.wav"
+            ]
+        }
+    },
+    {
+        "id": 11122,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003225.wav"
+            ]
+        }
+    },
+    {
+        "id": 11136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002866.wav"
+            ]
+        }
+    },
+    {
+        "id": 11139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001193.wav"
+            ]
+        }
+    },
+    {
+        "id": 11142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001028.wav"
+            ]
+        }
+    },
+    {
+        "id": 11144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006256.wav"
+            ]
+        }
+    },
+    {
+        "id": 11153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007099.wav"
+            ]
+        }
+    },
+    {
+        "id": 11156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007923.wav"
+            ]
+        }
+    },
+    {
+        "id": 11165,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006406.wav"
+            ]
+        }
+    },
+    {
+        "id": 11167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003679.wav"
+            ]
+        }
+    },
+    {
+        "id": 11169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003175.wav"
+            ]
+        }
+    },
+    {
+        "id": 11171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003496.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003496.wav"
+            ]
+        }
+    },
+    {
+        "id": 11182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004445.wav"
+            ]
+        }
+    },
+    {
+        "id": 11188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the drum on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000005.wav"
+            ]
+        }
+    },
+    {
+        "id": 11198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008181.wav"
+            ]
+        }
+    },
+    {
+        "id": 11210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005049.wav"
+            ]
+        }
+    },
+    {
+        "id": 11211,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003146.wav"
+            ]
+        }
+    },
+    {
+        "id": 11212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006175.wav"
+            ]
+        }
+    },
+    {
+        "id": 11213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003002.wav"
+            ]
+        }
+    },
+    {
+        "id": 11215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004124.wav"
+            ]
+        }
+    },
+    {
+        "id": 11217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone on the right more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001462.wav"
+            ]
+        }
+    },
+    {
+        "id": 11220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002938.wav"
+            ]
+        }
+    },
+    {
+        "id": 11227,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006478.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006478.wav"
+            ]
+        }
+    },
+    {
+        "id": 11229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001274.wav"
+            ]
+        }
+    },
+    {
+        "id": 11232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005156.wav"
+            ]
+        }
+    },
+    {
+        "id": 11242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004270.wav"
+            ]
+        }
+    },
+    {
+        "id": 11244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006516.wav"
+            ]
+        }
+    },
+    {
+        "id": 11262,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007927.wav"
+            ]
+        }
+    },
+    {
+        "id": 11264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005992.wav"
+            ]
+        }
+    },
+    {
+        "id": 11266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006516.wav"
+            ]
+        }
+    },
+    {
+        "id": 11270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005629.wav"
+            ]
+        }
+    },
+    {
+        "id": 11271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003939.wav"
+            ]
+        }
+    },
+    {
+        "id": 11283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008702.wav"
+            ]
+        }
+    },
+    {
+        "id": 11289,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006989.wav"
+            ]
+        }
+    },
+    {
+        "id": 11293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000411.wav"
+            ]
+        }
+    },
+    {
+        "id": 11297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000043.wav"
+            ]
+        }
+    },
+    {
+        "id": 11302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007945.wav"
+            ]
+        }
+    },
+    {
+        "id": 11308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003175.wav"
+            ]
+        }
+    },
+    {
+        "id": 11311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008840.wav"
+            ]
+        }
+    },
+    {
+        "id": 11321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004503.wav"
+            ]
+        }
+    },
+    {
+        "id": 11327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005960.wav"
+            ]
+        }
+    },
+    {
+        "id": 11330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001833.wav"
+            ]
+        }
+    },
+    {
+        "id": 11339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003829.wav"
+            ]
+        }
+    },
+    {
+        "id": 11342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002612.wav"
+            ]
+        }
+    },
+    {
+        "id": 11349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002855.wav"
+            ]
+        }
+    },
+    {
+        "id": 11352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007824.wav"
+            ]
+        }
+    },
+    {
+        "id": 11354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002791.wav"
+            ]
+        }
+    },
+    {
+        "id": 11365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001488.wav"
+            ]
+        }
+    },
+    {
+        "id": 11377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008779.wav"
+            ]
+        }
+    },
+    {
+        "id": 11380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000143.wav"
+            ]
+        }
+    },
+    {
+        "id": 11397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008472.wav"
+            ]
+        }
+    },
+    {
+        "id": 11406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006612.wav"
+            ]
+        }
+    },
+    {
+        "id": 11409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008730.wav"
+            ]
+        }
+    },
+    {
+        "id": 11417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002005.wav"
+            ]
+        }
+    },
+    {
+        "id": 11418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the violin on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001064.wav"
+            ]
+        }
+    },
+    {
+        "id": 11422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008774.wav"
+            ]
+        }
+    },
+    {
+        "id": 11427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005557.wav"
+            ]
+        }
+    },
+    {
+        "id": 11432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005813.wav"
+            ]
+        }
+    },
+    {
+        "id": 11436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002823.wav"
+            ]
+        }
+    },
+    {
+        "id": 11438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007731.wav"
+            ]
+        }
+    },
+    {
+        "id": 11448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right more rhythmic than the piano on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000835.wav"
+            ]
+        }
+    },
+    {
+        "id": 11455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the middle more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005082.wav"
+            ]
+        }
+    },
+    {
+        "id": 11460,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003493.wav"
+            ]
+        }
+    },
+    {
+        "id": 11476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000159.wav"
+            ]
+        }
+    },
+    {
+        "id": 11478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001101.wav"
+            ]
+        }
+    },
+    {
+        "id": 11492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001808.wav"
+            ]
+        }
+    },
+    {
+        "id": 11493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008214.wav"
+            ]
+        }
+    },
+    {
+        "id": 11506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002035.wav"
+            ]
+        }
+    },
+    {
+        "id": 11517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007263.wav"
+            ]
+        }
+    },
+    {
+        "id": 11518,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004495.wav"
+            ]
+        }
+    },
+    {
+        "id": 11523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008040.wav"
+            ]
+        }
+    },
+    {
+        "id": 11524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000947.wav"
+            ]
+        }
+    },
+    {
+        "id": 11526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass on the left louder than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004662.wav"
+            ]
+        }
+    },
+    {
+        "id": 11527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001846.wav"
+            ]
+        }
+    },
+    {
+        "id": 11537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007706.wav"
+            ]
+        }
+    },
+    {
+        "id": 11547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004388.wav"
+            ]
+        }
+    },
+    {
+        "id": 11552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008748.wav"
+            ]
+        }
+    },
+    {
+        "id": 11553,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008753.wav"
+            ]
+        }
+    },
+    {
+        "id": 11586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003203.wav"
+            ]
+        }
+    },
+    {
+        "id": 11588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006243.wav"
+            ]
+        }
+    },
+    {
+        "id": 11590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002925.wav"
+            ]
+        }
+    },
+    {
+        "id": 11601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004852.wav"
+            ]
+        }
+    },
+    {
+        "id": 11603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001781.wav"
+            ]
+        }
+    },
+    {
+        "id": 11612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003835.wav"
+            ]
+        }
+    },
+    {
+        "id": 11618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008333.wav"
+            ]
+        }
+    },
+    {
+        "id": 11622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005610.wav"
+            ]
+        }
+    },
+    {
+        "id": 11635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007883.wav"
+            ]
+        }
+    },
+    {
+        "id": 11637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000376.wav"
+            ]
+        }
+    },
+    {
+        "id": 11640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001581.wav"
+            ]
+        }
+    },
+    {
+        "id": 11646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005629.wav"
+            ]
+        }
+    },
+    {
+        "id": 11651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005857.wav"
+            ]
+        }
+    },
+    {
+        "id": 11656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008123.wav"
+            ]
+        }
+    },
+    {
+        "id": 11673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004243.wav"
+            ]
+        }
+    },
+    {
+        "id": 11678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007370.wav"
+            ]
+        }
+    },
+    {
+        "id": 11681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005188.wav"
+            ]
+        }
+    },
+    {
+        "id": 11688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008152.wav"
+            ]
+        }
+    },
+    {
+        "id": 11691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004795.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004795.wav"
+            ]
+        }
+    },
+    {
+        "id": 11694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003569.wav"
+            ]
+        }
+    },
+    {
+        "id": 11696,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008842.wav"
+            ]
+        }
+    },
+    {
+        "id": 11704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004193.wav"
+            ]
+        }
+    },
+    {
+        "id": 11709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002917.wav"
+            ]
+        }
+    },
+    {
+        "id": 11710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005697.wav"
+            ]
+        }
+    },
+    {
+        "id": 11712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003373.wav"
+            ]
+        }
+    },
+    {
+        "id": 11714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003411.wav"
+            ]
+        }
+    },
+    {
+        "id": 11720,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007346.wav"
+            ]
+        }
+    },
+    {
+        "id": 11721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001968.wav"
+            ]
+        }
+    },
+    {
+        "id": 11737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007304.wav"
+            ]
+        }
+    },
+    {
+        "id": 11743,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001067.wav"
+            ]
+        }
+    },
+    {
+        "id": 11755,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006129.wav"
+            ]
+        }
+    },
+    {
+        "id": 11763,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe on the right louder than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004102.wav"
+            ]
+        }
+    },
+    {
+        "id": 11772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004643.wav"
+            ]
+        }
+    },
+    {
+        "id": 11774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004653.wav"
+            ]
+        }
+    },
+    {
+        "id": 11775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002595.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002595.wav"
+            ]
+        }
+    },
+    {
+        "id": 11778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001501.wav"
+            ]
+        }
+    },
+    {
+        "id": 11784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002230.wav"
+            ]
+        }
+    },
+    {
+        "id": 11787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003630.wav"
+            ]
+        }
+    },
+    {
+        "id": 11789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008942.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008942.wav"
+            ]
+        }
+    },
+    {
+        "id": 11790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005820.wav"
+            ]
+        }
+    },
+    {
+        "id": 11791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000427.wav"
+            ]
+        }
+    },
+    {
+        "id": 11795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006011.wav"
+            ]
+        }
+    },
+    {
+        "id": 11807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007461.wav"
+            ]
+        }
+    },
+    {
+        "id": 11809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006485.wav"
+            ]
+        }
+    },
+    {
+        "id": 11812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003489.wav"
+            ]
+        }
+    },
+    {
+        "id": 11815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006147.wav"
+            ]
+        }
+    },
+    {
+        "id": 11820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006812.wav"
+            ]
+        }
+    },
+    {
+        "id": 11824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008664.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008664.wav"
+            ]
+        }
+    },
+    {
+        "id": 11825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007572.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007572.wav"
+            ]
+        }
+    },
+    {
+        "id": 11829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004556.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004556.wav"
+            ]
+        }
+    },
+    {
+        "id": 11842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001113.wav"
+            ]
+        }
+    },
+    {
+        "id": 11843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003752.wav"
+            ]
+        }
+    },
+    {
+        "id": 11848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006394.wav"
+            ]
+        }
+    },
+    {
+        "id": 11854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008124.wav"
+            ]
+        }
+    },
+    {
+        "id": 11855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008443.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008443.wav"
+            ]
+        }
+    },
+    {
+        "id": 11858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006532.wav"
+            ]
+        }
+    },
+    {
+        "id": 11861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004209.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004209.wav"
+            ]
+        }
+    },
+    {
+        "id": 11864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003413.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003413.wav"
+            ]
+        }
+    },
+    {
+        "id": 11867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001904.wav"
+            ]
+        }
+    },
+    {
+        "id": 11882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006057.wav"
+            ]
+        }
+    },
+    {
+        "id": 11901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003291.wav"
+            ]
+        }
+    },
+    {
+        "id": 11902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005893.wav"
+            ]
+        }
+    },
+    {
+        "id": 11906,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002962.wav"
+            ]
+        }
+    },
+    {
+        "id": 11912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003462.wav"
+            ]
+        }
+    },
+    {
+        "id": 11913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004560.wav"
+            ]
+        }
+    },
+    {
+        "id": 11914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006176.wav"
+            ]
+        }
+    },
+    {
+        "id": 11916,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001919.wav"
+            ]
+        }
+    },
+    {
+        "id": 11924,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001736.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001736.wav"
+            ]
+        }
+    },
+    {
+        "id": 11930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003015.wav"
+            ]
+        }
+    },
+    {
+        "id": 11931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000908.wav"
+            ]
+        }
+    },
+    {
+        "id": 11932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002608.wav"
+            ]
+        }
+    },
+    {
+        "id": 11933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005542.wav"
+            ]
+        }
+    },
+    {
+        "id": 11936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008272.wav"
+            ]
+        }
+    },
+    {
+        "id": 11938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich electric_bass makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002821.wav"
+            ]
+        }
+    },
+    {
+        "id": 11939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left louder than the banjo on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006171.wav"
+            ]
+        }
+    },
+    {
+        "id": 11940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006536.wav"
+            ]
+        }
+    },
+    {
+        "id": 11943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002861.wav"
+            ]
+        }
+    },
+    {
+        "id": 11949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002171.wav"
+            ]
+        }
+    },
+    {
+        "id": 11958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003921.wav"
+            ]
+        }
+    },
+    {
+        "id": 11963,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003954.wav"
+            ]
+        }
+    },
+    {
+        "id": 11977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003693.wav"
+            ]
+        }
+    },
+    {
+        "id": 11978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005274.wav"
+            ]
+        }
+    },
+    {
+        "id": 11981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004554.wav"
+            ]
+        }
+    },
+    {
+        "id": 11986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001968.wav"
+            ]
+        }
+    },
+    {
+        "id": 11999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001503.wav"
+            ]
+        }
+    },
+    {
+        "id": 12001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007934.wav"
+            ]
+        }
+    },
+    {
+        "id": 12004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001350.wav"
+            ]
+        }
+    },
+    {
+        "id": 12010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002812.wav"
+            ]
+        }
+    },
+    {
+        "id": 12013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005702.wav"
+            ]
+        }
+    },
+    {
+        "id": 12016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006386.wav"
+            ]
+        }
+    },
+    {
+        "id": 12017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001291.wav"
+            ]
+        }
+    },
+    {
+        "id": 12021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the left more rhythmic than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008692.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008692.wav"
+            ]
+        }
+    },
+    {
+        "id": 12023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006163.wav"
+            ]
+        }
+    },
+    {
+        "id": 12027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004253.wav"
+            ]
+        }
+    },
+    {
+        "id": 12044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007436.wav"
+            ]
+        }
+    },
+    {
+        "id": 12050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002621.wav"
+            ]
+        }
+    },
+    {
+        "id": 12051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006106.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006106.wav"
+            ]
+        }
+    },
+    {
+        "id": 12052,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right more rhythmic than the bassoon on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002781.wav"
+            ]
+        }
+    },
+    {
+        "id": 12059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005951.wav"
+            ]
+        }
+    },
+    {
+        "id": 12069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006596.wav"
+            ]
+        }
+    },
+    {
+        "id": 12073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008816.wav"
+            ]
+        }
+    },
+    {
+        "id": 12075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005515.wav"
+            ]
+        }
+    },
+    {
+        "id": 12080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000324.wav"
+            ]
+        }
+    },
+    {
+        "id": 12084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008655.wav"
+            ]
+        }
+    },
+    {
+        "id": 12089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001202.wav"
+            ]
+        }
+    },
+    {
+        "id": 12090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001312.wav"
+            ]
+        }
+    },
+    {
+        "id": 12100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000230.wav"
+            ]
+        }
+    },
+    {
+        "id": 12105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005159.wav"
+            ]
+        }
+    },
+    {
+        "id": 12121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002655.wav"
+            ]
+        }
+    },
+    {
+        "id": 12123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005546.wav"
+            ]
+        }
+    },
+    {
+        "id": 12137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006175.wav"
+            ]
+        }
+    },
+    {
+        "id": 12150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002775.wav"
+            ]
+        }
+    },
+    {
+        "id": 12152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001834.wav"
+            ]
+        }
+    },
+    {
+        "id": 12155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and pipa instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001981.wav"
+            ]
+        }
+    },
+    {
+        "id": 12163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008664.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008664.wav"
+            ]
+        }
+    },
+    {
+        "id": 12171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000743.wav"
+            ]
+        }
+    },
+    {
+        "id": 12172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005198.wav"
+            ]
+        }
+    },
+    {
+        "id": 12173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000405.wav"
+            ]
+        }
+    },
+    {
+        "id": 12177,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000023.wav"
+            ]
+        }
+    },
+    {
+        "id": 12180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002672.wav"
+            ]
+        }
+    },
+    {
+        "id": 12214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008069.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008069.wav"
+            ]
+        }
+    },
+    {
+        "id": 12217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left more rhythmic than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006835.wav"
+            ]
+        }
+    },
+    {
+        "id": 12219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005582.wav"
+            ]
+        }
+    },
+    {
+        "id": 12221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006576.wav"
+            ]
+        }
+    },
+    {
+        "id": 12223,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005366.wav"
+            ]
+        }
+    },
+    {
+        "id": 12224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001393.wav"
+            ]
+        }
+    },
+    {
+        "id": 12226,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006098.wav"
+            ]
+        }
+    },
+    {
+        "id": 12234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000439.wav"
+            ]
+        }
+    },
+    {
+        "id": 12236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005416.wav"
+            ]
+        }
+    },
+    {
+        "id": 12241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006411.wav"
+            ]
+        }
+    },
+    {
+        "id": 12243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006131.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006131.wav"
+            ]
+        }
+    },
+    {
+        "id": 12249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006022.wav"
+            ]
+        }
+    },
+    {
+        "id": 12253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008152.wav"
+            ]
+        }
+    },
+    {
+        "id": 12254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000923.wav"
+            ]
+        }
+    },
+    {
+        "id": 12256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006147.wav"
+            ]
+        }
+    },
+    {
+        "id": 12258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002341.wav"
+            ]
+        }
+    },
+    {
+        "id": 12259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000260.wav"
+            ]
+        }
+    },
+    {
+        "id": 12290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001401.wav"
+            ]
+        }
+    },
+    {
+        "id": 12294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003116.wav"
+            ]
+        }
+    },
+    {
+        "id": 12297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000416.wav"
+            ]
+        }
+    },
+    {
+        "id": 12301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005287.wav"
+            ]
+        }
+    },
+    {
+        "id": 12315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004424.wav"
+            ]
+        }
+    },
+    {
+        "id": 12319,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000297.wav"
+            ]
+        }
+    },
+    {
+        "id": 12325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007456.wav"
+            ]
+        }
+    },
+    {
+        "id": 12327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008952.wav"
+            ]
+        }
+    },
+    {
+        "id": 12330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006876.wav"
+            ]
+        }
+    },
+    {
+        "id": 12332,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006161.wav"
+            ]
+        }
+    },
+    {
+        "id": 12333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008100.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008100.wav"
+            ]
+        }
+    },
+    {
+        "id": 12334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007195.wav"
+            ]
+        }
+    },
+    {
+        "id": 12337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001348.wav"
+            ]
+        }
+    },
+    {
+        "id": 12338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000863.wav"
+            ]
+        }
+    },
+    {
+        "id": 12339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000174.wav"
+            ]
+        }
+    },
+    {
+        "id": 12345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002476.wav"
+            ]
+        }
+    },
+    {
+        "id": 12346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003788.wav"
+            ]
+        }
+    },
+    {
+        "id": 12354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004718.wav"
+            ]
+        }
+    },
+    {
+        "id": 12360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007109.wav"
+            ]
+        }
+    },
+    {
+        "id": 12363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001806.wav"
+            ]
+        }
+    },
+    {
+        "id": 12367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001637.wav"
+            ]
+        }
+    },
+    {
+        "id": 12370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008405.wav"
+            ]
+        }
+    },
+    {
+        "id": 12371,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004838.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004838.wav"
+            ]
+        }
+    },
+    {
+        "id": 12387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left louder than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002892.wav"
+            ]
+        }
+    },
+    {
+        "id": 12390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008131.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008131.wav"
+            ]
+        }
+    },
+    {
+        "id": 12392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003939.wav"
+            ]
+        }
+    },
+    {
+        "id": 12395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001090.wav"
+            ]
+        }
+    },
+    {
+        "id": 12401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004874.wav"
+            ]
+        }
+    },
+    {
+        "id": 12410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001029.wav"
+            ]
+        }
+    },
+    {
+        "id": 12411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005014.wav"
+            ]
+        }
+    },
+    {
+        "id": 12413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005704.wav"
+            ]
+        }
+    },
+    {
+        "id": 12419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002501.wav"
+            ]
+        }
+    },
+    {
+        "id": 12437,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007452.wav"
+            ]
+        }
+    },
+    {
+        "id": 12440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003520.wav"
+            ]
+        }
+    },
+    {
+        "id": 12444,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005262.wav"
+            ]
+        }
+    },
+    {
+        "id": 12448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007954.wav"
+            ]
+        }
+    },
+    {
+        "id": 12453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005864.wav"
+            ]
+        }
+    },
+    {
+        "id": 12462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007119.wav"
+            ]
+        }
+    },
+    {
+        "id": 12466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002716.wav"
+            ]
+        }
+    },
+    {
+        "id": 12467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000669.wav"
+            ]
+        }
+    },
+    {
+        "id": 12468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right louder than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005588.wav"
+            ]
+        }
+    },
+    {
+        "id": 12469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004112.wav"
+            ]
+        }
+    },
+    {
+        "id": 12470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003573.wav"
+            ]
+        }
+    },
+    {
+        "id": 12476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003739.wav"
+            ]
+        }
+    },
+    {
+        "id": 12483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006851.wav"
+            ]
+        }
+    },
+    {
+        "id": 12488,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000458.wav"
+            ]
+        }
+    },
+    {
+        "id": 12491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006831.wav"
+            ]
+        }
+    },
+    {
+        "id": 12502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008439.wav"
+            ]
+        }
+    },
+    {
+        "id": 12513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002054.wav"
+            ]
+        }
+    },
+    {
+        "id": 12515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003348.wav"
+            ]
+        }
+    },
+    {
+        "id": 12517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002188.wav"
+            ]
+        }
+    },
+    {
+        "id": 12522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002710.wav"
+            ]
+        }
+    },
+    {
+        "id": 12525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the middle more rhythmic than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002327.wav"
+            ]
+        }
+    },
+    {
+        "id": 12532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008814.wav"
+            ]
+        }
+    },
+    {
+        "id": 12541,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001358.wav"
+            ]
+        }
+    },
+    {
+        "id": 12547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006723.wav"
+            ]
+        }
+    },
+    {
+        "id": 12550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003994.wav"
+            ]
+        }
+    },
+    {
+        "id": 12554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001237.wav"
+            ]
+        }
+    },
+    {
+        "id": 12555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001219.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001219.wav"
+            ]
+        }
+    },
+    {
+        "id": 12563,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004723.wav"
+            ]
+        }
+    },
+    {
+        "id": 12568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006777.wav"
+            ]
+        }
+    },
+    {
+        "id": 12569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002542.wav"
+            ]
+        }
+    },
+    {
+        "id": 12572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002762.wav"
+            ]
+        }
+    },
+    {
+        "id": 12574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000993.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000993.wav"
+            ]
+        }
+    },
+    {
+        "id": 12581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001587.wav"
+            ]
+        }
+    },
+    {
+        "id": 12587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008487.wav"
+            ]
+        }
+    },
+    {
+        "id": 12588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006342.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006342.wav"
+            ]
+        }
+    },
+    {
+        "id": 12594,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001467.wav"
+            ]
+        }
+    },
+    {
+        "id": 12596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003562.wav"
+            ]
+        }
+    },
+    {
+        "id": 12610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004780.wav"
+            ]
+        }
+    },
+    {
+        "id": 12617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003699.wav"
+            ]
+        }
+    },
+    {
+        "id": 12622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004298.wav"
+            ]
+        }
+    },
+    {
+        "id": 12625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006345.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006345.wav"
+            ]
+        }
+    },
+    {
+        "id": 12627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007488.wav"
+            ]
+        }
+    },
+    {
+        "id": 12635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003124.wav"
+            ]
+        }
+    },
+    {
+        "id": 12638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002362.wav"
+            ]
+        }
+    },
+    {
+        "id": 12645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002523.wav"
+            ]
+        }
+    },
+    {
+        "id": 12650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001085.wav"
+            ]
+        }
+    },
+    {
+        "id": 12651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008142.wav"
+            ]
+        }
+    },
+    {
+        "id": 12652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001848.wav"
+            ]
+        }
+    },
+    {
+        "id": 12659,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005811.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005811.wav"
+            ]
+        }
+    },
+    {
+        "id": 12665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006471.wav"
+            ]
+        }
+    },
+    {
+        "id": 12668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001064.wav"
+            ]
+        }
+    },
+    {
+        "id": 12669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002090.wav"
+            ]
+        }
+    },
+    {
+        "id": 12677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002314.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002314.wav"
+            ]
+        }
+    },
+    {
+        "id": 12682,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002721.wav"
+            ]
+        }
+    },
+    {
+        "id": 12684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000304.wav"
+            ]
+        }
+    },
+    {
+        "id": 12691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004562.wav"
+            ]
+        }
+    },
+    {
+        "id": 12695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008545.wav"
+            ]
+        }
+    },
+    {
+        "id": 12696,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008007.wav"
+            ]
+        }
+    },
+    {
+        "id": 12705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005636.wav"
+            ]
+        }
+    },
+    {
+        "id": 12712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006202.wav"
+            ]
+        }
+    },
+    {
+        "id": 12713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008496.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008496.wav"
+            ]
+        }
+    },
+    {
+        "id": 12716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005045.wav"
+            ]
+        }
+    },
+    {
+        "id": 12717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008460.wav"
+            ]
+        }
+    },
+    {
+        "id": 12718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001792.wav"
+            ]
+        }
+    },
+    {
+        "id": 12720,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006065.wav"
+            ]
+        }
+    },
+    {
+        "id": 12727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008511.wav"
+            ]
+        }
+    },
+    {
+        "id": 12731,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005559.wav"
+            ]
+        }
+    },
+    {
+        "id": 12732,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004535.wav"
+            ]
+        }
+    },
+    {
+        "id": 12753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005601.wav"
+            ]
+        }
+    },
+    {
+        "id": 12757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005844.wav"
+            ]
+        }
+    },
+    {
+        "id": 12760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002803.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002803.wav"
+            ]
+        }
+    },
+    {
+        "id": 12761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003991.wav"
+            ]
+        }
+    },
+    {
+        "id": 12778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003111.wav"
+            ]
+        }
+    },
+    {
+        "id": 12784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002601.wav"
+            ]
+        }
+    },
+    {
+        "id": 12786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000464.wav"
+            ]
+        }
+    },
+    {
+        "id": 12794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006444.wav"
+            ]
+        }
+    },
+    {
+        "id": 12797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005404.wav"
+            ]
+        }
+    },
+    {
+        "id": 12826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000597.wav"
+            ]
+        }
+    },
+    {
+        "id": 12833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006619.wav"
+            ]
+        }
+    },
+    {
+        "id": 12838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005324.wav"
+            ]
+        }
+    },
+    {
+        "id": 12844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004821.wav"
+            ]
+        }
+    },
+    {
+        "id": 12847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002512.wav"
+            ]
+        }
+    },
+    {
+        "id": 12848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001406.wav"
+            ]
+        }
+    },
+    {
+        "id": 12849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002726.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002726.wav"
+            ]
+        }
+    },
+    {
+        "id": 12857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004855.wav"
+            ]
+        }
+    },
+    {
+        "id": 12860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008878.wav"
+            ]
+        }
+    },
+    {
+        "id": 12871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002985.wav"
+            ]
+        }
+    },
+    {
+        "id": 12874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003374.wav"
+            ]
+        }
+    },
+    {
+        "id": 12875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001790.wav"
+            ]
+        }
+    },
+    {
+        "id": 12884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007611.wav"
+            ]
+        }
+    },
+    {
+        "id": 12888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007052.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007052.wav"
+            ]
+        }
+    },
+    {
+        "id": 12891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007936.wav"
+            ]
+        }
+    },
+    {
+        "id": 12894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007608.wav"
+            ]
+        }
+    },
+    {
+        "id": 12895,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008647.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008647.wav"
+            ]
+        }
+    },
+    {
+        "id": 12896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002362.wav"
+            ]
+        }
+    },
+    {
+        "id": 12902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007960.wav"
+            ]
+        }
+    },
+    {
+        "id": 12904,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008644.wav"
+            ]
+        }
+    },
+    {
+        "id": 12909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000740.wav"
+            ]
+        }
+    },
+    {
+        "id": 12910,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008591.wav"
+            ]
+        }
+    },
+    {
+        "id": 12911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005651.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005651.wav"
+            ]
+        }
+    },
+    {
+        "id": 12912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006346.wav"
+            ]
+        }
+    },
+    {
+        "id": 12915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007198.wav"
+            ]
+        }
+    },
+    {
+        "id": 12920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000201.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000201.wav"
+            ]
+        }
+    },
+    {
+        "id": 12942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000280.wav"
+            ]
+        }
+    },
+    {
+        "id": 12947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007881.wav"
+            ]
+        }
+    },
+    {
+        "id": 12951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008261.wav"
+            ]
+        }
+    },
+    {
+        "id": 12952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004173.wav"
+            ]
+        }
+    },
+    {
+        "id": 12955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001773.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001773.wav"
+            ]
+        }
+    },
+    {
+        "id": 12960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008307.wav"
+            ]
+        }
+    },
+    {
+        "id": 12962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007047.wav"
+            ]
+        }
+    },
+    {
+        "id": 12965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005876.wav"
+            ]
+        }
+    },
+    {
+        "id": 12966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004272.wav"
+            ]
+        }
+    },
+    {
+        "id": 12970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004761.wav"
+            ]
+        }
+    },
+    {
+        "id": 12977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003864.wav"
+            ]
+        }
+    },
+    {
+        "id": 12980,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003837.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003837.wav"
+            ]
+        }
+    },
+    {
+        "id": 12988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007720.wav"
+            ]
+        }
+    },
+    {
+        "id": 12991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004691.wav"
+            ]
+        }
+    },
+    {
+        "id": 12998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005726.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005726.wav"
+            ]
+        }
+    },
+    {
+        "id": 13001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003408.wav"
+            ]
+        }
+    },
+    {
+        "id": 13003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006668.wav"
+            ]
+        }
+    },
+    {
+        "id": 13005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000377.wav"
+            ]
+        }
+    },
+    {
+        "id": 13006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005860.wav"
+            ]
+        }
+    },
+    {
+        "id": 13011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001654.wav"
+            ]
+        }
+    },
+    {
+        "id": 13014,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001234.wav"
+            ]
+        }
+    },
+    {
+        "id": 13018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008836.wav"
+            ]
+        }
+    },
+    {
+        "id": 13027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006149.wav"
+            ]
+        }
+    },
+    {
+        "id": 13028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003266.wav"
+            ]
+        }
+    },
+    {
+        "id": 13030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002795.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002795.wav"
+            ]
+        }
+    },
+    {
+        "id": 13032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008919.wav"
+            ]
+        }
+    },
+    {
+        "id": 13034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007447.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007447.wav"
+            ]
+        }
+    },
+    {
+        "id": 13040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005232.wav"
+            ]
+        }
+    },
+    {
+        "id": 13045,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007674.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007674.wav"
+            ]
+        }
+    },
+    {
+        "id": 13051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006581.wav"
+            ]
+        }
+    },
+    {
+        "id": 13053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003694.wav"
+            ]
+        }
+    },
+    {
+        "id": 13056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002448.wav"
+            ]
+        }
+    },
+    {
+        "id": 13065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000725.wav"
+            ]
+        }
+    },
+    {
+        "id": 13068,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008258.wav"
+            ]
+        }
+    },
+    {
+        "id": 13070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005268.wav"
+            ]
+        }
+    },
+    {
+        "id": 13074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001798.wav"
+            ]
+        }
+    },
+    {
+        "id": 13079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006087.wav"
+            ]
+        }
+    },
+    {
+        "id": 13082,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004206.wav"
+            ]
+        }
+    },
+    {
+        "id": 13085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001841.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001841.wav"
+            ]
+        }
+    },
+    {
+        "id": 13087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000519.wav"
+            ]
+        }
+    },
+    {
+        "id": 13089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003908.wav"
+            ]
+        }
+    },
+    {
+        "id": 13090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006135.wav"
+            ]
+        }
+    },
+    {
+        "id": 13093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004802.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004802.wav"
+            ]
+        }
+    },
+    {
+        "id": 13096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006874.wav"
+            ]
+        }
+    },
+    {
+        "id": 13097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001573.wav"
+            ]
+        }
+    },
+    {
+        "id": 13098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001632.wav"
+            ]
+        }
+    },
+    {
+        "id": 13099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007912.wav"
+            ]
+        }
+    },
+    {
+        "id": 13101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001787.wav"
+            ]
+        }
+    },
+    {
+        "id": 13108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003355.wav"
+            ]
+        }
+    },
+    {
+        "id": 13110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001861.wav"
+            ]
+        }
+    },
+    {
+        "id": 13113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005771.wav"
+            ]
+        }
+    },
+    {
+        "id": 13126,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007188.wav"
+            ]
+        }
+    },
+    {
+        "id": 13129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001916.wav"
+            ]
+        }
+    },
+    {
+        "id": 13133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and xylophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001462.wav"
+            ]
+        }
+    },
+    {
+        "id": 13138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000949.wav"
+            ]
+        }
+    },
+    {
+        "id": 13142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007698.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007698.wav"
+            ]
+        }
+    },
+    {
+        "id": 13143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001316.wav"
+            ]
+        }
+    },
+    {
+        "id": 13144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006408.wav"
+            ]
+        }
+    },
+    {
+        "id": 13145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007550.wav"
+            ]
+        }
+    },
+    {
+        "id": 13154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004477.wav"
+            ]
+        }
+    },
+    {
+        "id": 13158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003260.wav"
+            ]
+        }
+    },
+    {
+        "id": 13159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003314.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003314.wav"
+            ]
+        }
+    },
+    {
+        "id": 13160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002364.wav"
+            ]
+        }
+    },
+    {
+        "id": 13163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001547.wav"
+            ]
+        }
+    },
+    {
+        "id": 13167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002534.wav"
+            ]
+        }
+    },
+    {
+        "id": 13169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005486.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005486.wav"
+            ]
+        }
+    },
+    {
+        "id": 13170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000164.wav"
+            ]
+        }
+    },
+    {
+        "id": 13177,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004532.wav"
+            ]
+        }
+    },
+    {
+        "id": 13179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001111.wav"
+            ]
+        }
+    },
+    {
+        "id": 13181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000186.wav"
+            ]
+        }
+    },
+    {
+        "id": 13192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001959.wav"
+            ]
+        }
+    },
+    {
+        "id": 13194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006842.wav"
+            ]
+        }
+    },
+    {
+        "id": 13195,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008865.wav"
+            ]
+        }
+    },
+    {
+        "id": 13198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the right more rhythmic than the guzheng on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005297.wav"
+            ]
+        }
+    },
+    {
+        "id": 13206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007805.wav"
+            ]
+        }
+    },
+    {
+        "id": 13211,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007035.wav"
+            ]
+        }
+    },
+    {
+        "id": 13216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004998.wav"
+            ]
+        }
+    },
+    {
+        "id": 13238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003541.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003541.wav"
+            ]
+        }
+    },
+    {
+        "id": 13241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006890.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006890.wav"
+            ]
+        }
+    },
+    {
+        "id": 13245,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000576.wav"
+            ]
+        }
+    },
+    {
+        "id": 13251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000065.wav"
+            ]
+        }
+    },
+    {
+        "id": 13256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003438.wav"
+            ]
+        }
+    },
+    {
+        "id": 13258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001708.wav"
+            ]
+        }
+    },
+    {
+        "id": 13260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002119.wav"
+            ]
+        }
+    },
+    {
+        "id": 13269,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001181.wav"
+            ]
+        }
+    },
+    {
+        "id": 13271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007002.wav"
+            ]
+        }
+    },
+    {
+        "id": 13277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003676.wav"
+            ]
+        }
+    },
+    {
+        "id": 13280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002566.wav"
+            ]
+        }
+    },
+    {
+        "id": 13283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005748.wav"
+            ]
+        }
+    },
+    {
+        "id": 13290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008945.wav"
+            ]
+        }
+    },
+    {
+        "id": 13293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005389.wav"
+            ]
+        }
+    },
+    {
+        "id": 13298,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008910.wav"
+            ]
+        }
+    },
+    {
+        "id": 13303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007120.wav"
+            ]
+        }
+    },
+    {
+        "id": 13309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005196.wav"
+            ]
+        }
+    },
+    {
+        "id": 13312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004177.wav"
+            ]
+        }
+    },
+    {
+        "id": 13313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006374.wav"
+            ]
+        }
+    },
+    {
+        "id": 13314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003954.wav"
+            ]
+        }
+    },
+    {
+        "id": 13321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001501.wav"
+            ]
+        }
+    },
+    {
+        "id": 13324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004160.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004160.wav"
+            ]
+        }
+    },
+    {
+        "id": 13333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003698.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003698.wav"
+            ]
+        }
+    },
+    {
+        "id": 13336,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000735.wav"
+            ]
+        }
+    },
+    {
+        "id": 13337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003628.wav"
+            ]
+        }
+    },
+    {
+        "id": 13340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003885.wav"
+            ]
+        }
+    },
+    {
+        "id": 13342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005653.wav"
+            ]
+        }
+    },
+    {
+        "id": 13347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003650.wav"
+            ]
+        }
+    },
+    {
+        "id": 13355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005323.wav"
+            ]
+        }
+    },
+    {
+        "id": 13364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001524.wav"
+            ]
+        }
+    },
+    {
+        "id": 13370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006320.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006320.wav"
+            ]
+        }
+    },
+    {
+        "id": 13380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007038.wav"
+            ]
+        }
+    },
+    {
+        "id": 13389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004164.wav"
+            ]
+        }
+    },
+    {
+        "id": 13396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006282.wav"
+            ]
+        }
+    },
+    {
+        "id": 13400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000098.wav"
+            ]
+        }
+    },
+    {
+        "id": 13401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002754.wav"
+            ]
+        }
+    },
+    {
+        "id": 13405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007076.wav"
+            ]
+        }
+    },
+    {
+        "id": 13406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001809.wav"
+            ]
+        }
+    },
+    {
+        "id": 13413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002512.wav"
+            ]
+        }
+    },
+    {
+        "id": 13417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007477.wav"
+            ]
+        }
+    },
+    {
+        "id": 13418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000431.wav"
+            ]
+        }
+    },
+    {
+        "id": 13422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005695.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005695.wav"
+            ]
+        }
+    },
+    {
+        "id": 13426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002962.wav"
+            ]
+        }
+    },
+    {
+        "id": 13440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008480.wav"
+            ]
+        }
+    },
+    {
+        "id": 13441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003471.wav"
+            ]
+        }
+    },
+    {
+        "id": 13445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008697.wav"
+            ]
+        }
+    },
+    {
+        "id": 13456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000986.wav"
+            ]
+        }
+    },
+    {
+        "id": 13464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001793.wav"
+            ]
+        }
+    },
+    {
+        "id": 13471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005214.wav"
+            ]
+        }
+    },
+    {
+        "id": 13476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002233.wav"
+            ]
+        }
+    },
+    {
+        "id": 13481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001685.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001685.wav"
+            ]
+        }
+    },
+    {
+        "id": 13495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003416.wav"
+            ]
+        }
+    },
+    {
+        "id": 13498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008627.wav"
+            ]
+        }
+    },
+    {
+        "id": 13502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004670.wav"
+            ]
+        }
+    },
+    {
+        "id": 13510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003876.wav"
+            ]
+        }
+    },
+    {
+        "id": 13511,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001025.wav"
+            ]
+        }
+    },
+    {
+        "id": 13513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007568.wav"
+            ]
+        }
+    },
+    {
+        "id": 13515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001472.wav"
+            ]
+        }
+    },
+    {
+        "id": 13525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005613.wav"
+            ]
+        }
+    },
+    {
+        "id": 13531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008175.wav"
+            ]
+        }
+    },
+    {
+        "id": 13539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000232.wav"
+            ]
+        }
+    },
+    {
+        "id": 13550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007923.wav"
+            ]
+        }
+    },
+    {
+        "id": 13564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008322.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008322.wav"
+            ]
+        }
+    },
+    {
+        "id": 13565,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001697.wav"
+            ]
+        }
+    },
+    {
+        "id": 13575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007473.wav"
+            ]
+        }
+    },
+    {
+        "id": 13577,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000885.wav"
+            ]
+        }
+    },
+    {
+        "id": 13578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000751.wav"
+            ]
+        }
+    },
+    {
+        "id": 13589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007173.wav"
+            ]
+        }
+    },
+    {
+        "id": 13592,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004904.wav"
+            ]
+        }
+    },
+    {
+        "id": 13593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005966.wav"
+            ]
+        }
+    },
+    {
+        "id": 13596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008220.wav"
+            ]
+        }
+    },
+    {
+        "id": 13598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007355.wav"
+            ]
+        }
+    },
+    {
+        "id": 13603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004085.wav"
+            ]
+        }
+    },
+    {
+        "id": 13604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005840.wav"
+            ]
+        }
+    },
+    {
+        "id": 13609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002380.wav"
+            ]
+        }
+    },
+    {
+        "id": 13610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007014.wav"
+            ]
+        }
+    },
+    {
+        "id": 13618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007281.wav"
+            ]
+        }
+    },
+    {
+        "id": 13619,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006499.wav"
+            ]
+        }
+    },
+    {
+        "id": 13620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002473.wav"
+            ]
+        }
+    },
+    {
+        "id": 13630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000060.wav"
+            ]
+        }
+    },
+    {
+        "id": 13633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008148.wav"
+            ]
+        }
+    },
+    {
+        "id": 13636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008337.wav"
+            ]
+        }
+    },
+    {
+        "id": 13639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001901.wav"
+            ]
+        }
+    },
+    {
+        "id": 13650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008225.wav"
+            ]
+        }
+    },
+    {
+        "id": 13653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004579.wav"
+            ]
+        }
+    },
+    {
+        "id": 13655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005142.wav"
+            ]
+        }
+    },
+    {
+        "id": 13661,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000717.wav"
+            ]
+        }
+    },
+    {
+        "id": 13664,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007113.wav"
+            ]
+        }
+    },
+    {
+        "id": 13669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006258.wav"
+            ]
+        }
+    },
+    {
+        "id": 13672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004972.wav"
+            ]
+        }
+    },
+    {
+        "id": 13678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002868.wav"
+            ]
+        }
+    },
+    {
+        "id": 13684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000989.wav"
+            ]
+        }
+    },
+    {
+        "id": 13685,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008308.wav"
+            ]
+        }
+    },
+    {
+        "id": 13691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008236.wav"
+            ]
+        }
+    },
+    {
+        "id": 13695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001197.wav"
+            ]
+        }
+    },
+    {
+        "id": 13696,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006429.wav"
+            ]
+        }
+    },
+    {
+        "id": 13702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001027.wav"
+            ]
+        }
+    },
+    {
+        "id": 13710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006313.wav"
+            ]
+        }
+    },
+    {
+        "id": 13715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008273.wav"
+            ]
+        }
+    },
+    {
+        "id": 13718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008196.wav"
+            ]
+        }
+    },
+    {
+        "id": 13738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005195.wav"
+            ]
+        }
+    },
+    {
+        "id": 13739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008724.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008724.wav"
+            ]
+        }
+    },
+    {
+        "id": 13740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004925.wav"
+            ]
+        }
+    },
+    {
+        "id": 13741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008974.wav"
+            ]
+        }
+    },
+    {
+        "id": 13742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006488.wav"
+            ]
+        }
+    },
+    {
+        "id": 13748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008548.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008548.wav"
+            ]
+        }
+    },
+    {
+        "id": 13749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007961.wav"
+            ]
+        }
+    },
+    {
+        "id": 13752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001506.wav"
+            ]
+        }
+    },
+    {
+        "id": 13758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003018.wav"
+            ]
+        }
+    },
+    {
+        "id": 13760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005007.wav"
+            ]
+        }
+    },
+    {
+        "id": 13766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007384.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007384.wav"
+            ]
+        }
+    },
+    {
+        "id": 13773,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003420.wav"
+            ]
+        }
+    },
+    {
+        "id": 13777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003818.wav"
+            ]
+        }
+    },
+    {
+        "id": 13781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000186.wav"
+            ]
+        }
+    },
+    {
+        "id": 13784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002745.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002745.wav"
+            ]
+        }
+    },
+    {
+        "id": 13799,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004823.wav"
+            ]
+        }
+    },
+    {
+        "id": 13816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000619.wav"
+            ]
+        }
+    },
+    {
+        "id": 13838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002417.wav"
+            ]
+        }
+    },
+    {
+        "id": 13841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002635.wav"
+            ]
+        }
+    },
+    {
+        "id": 13848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006623.wav"
+            ]
+        }
+    },
+    {
+        "id": 13859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002248.wav"
+            ]
+        }
+    },
+    {
+        "id": 13874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002463.wav"
+            ]
+        }
+    },
+    {
+        "id": 13876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003733.wav"
+            ]
+        }
+    },
+    {
+        "id": 13881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000398.wav"
+            ]
+        }
+    },
+    {
+        "id": 13907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001826.wav"
+            ]
+        }
+    },
+    {
+        "id": 13913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004151.wav"
+            ]
+        }
+    },
+    {
+        "id": 13919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008422.wav"
+            ]
+        }
+    },
+    {
+        "id": 13922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001453.wav"
+            ]
+        }
+    },
+    {
+        "id": 13929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005955.wav"
+            ]
+        }
+    },
+    {
+        "id": 13930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003206.wav"
+            ]
+        }
+    },
+    {
+        "id": 13937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000235.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000235.wav"
+            ]
+        }
+    },
+    {
+        "id": 13953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008415.wav"
+            ]
+        }
+    },
+    {
+        "id": 13957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007764.wav"
+            ]
+        }
+    },
+    {
+        "id": 13964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008931.wav"
+            ]
+        }
+    },
+    {
+        "id": 13967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003573.wav"
+            ]
+        }
+    },
+    {
+        "id": 13974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007013.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007013.wav"
+            ]
+        }
+    },
+    {
+        "id": 13976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008339.wav"
+            ]
+        }
+    },
+    {
+        "id": 13977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000909.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000909.wav"
+            ]
+        }
+    },
+    {
+        "id": 13981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003167.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003167.wav"
+            ]
+        }
+    },
+    {
+        "id": 13992,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008334.wav"
+            ]
+        }
+    },
+    {
+        "id": 13994,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001132.wav"
+            ]
+        }
+    },
+    {
+        "id": 13996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001988.wav"
+            ]
+        }
+    },
+    {
+        "id": 14003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002467.wav"
+            ]
+        }
+    },
+    {
+        "id": 14006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006681.wav"
+            ]
+        }
+    },
+    {
+        "id": 14010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001752.wav"
+            ]
+        }
+    },
+    {
+        "id": 14012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002351.wav"
+            ]
+        }
+    },
+    {
+        "id": 14028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000339.wav"
+            ]
+        }
+    },
+    {
+        "id": 14029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003559.wav"
+            ]
+        }
+    },
+    {
+        "id": 14032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003628.wav"
+            ]
+        }
+    },
+    {
+        "id": 14033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001642.wav"
+            ]
+        }
+    },
+    {
+        "id": 14042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002927.wav"
+            ]
+        }
+    },
+    {
+        "id": 14046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002054.wav"
+            ]
+        }
+    },
+    {
+        "id": 14048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007360.wav"
+            ]
+        }
+    },
+    {
+        "id": 14049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007868.wav"
+            ]
+        }
+    },
+    {
+        "id": 14051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008186.wav"
+            ]
+        }
+    },
+    {
+        "id": 14052,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001354.wav"
+            ]
+        }
+    },
+    {
+        "id": 14054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000923.wav"
+            ]
+        }
+    },
+    {
+        "id": 14056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003013.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003013.wav"
+            ]
+        }
+    },
+    {
+        "id": 14061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007090.wav"
+            ]
+        }
+    },
+    {
+        "id": 14064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001497.wav"
+            ]
+        }
+    },
+    {
+        "id": 14076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003475.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003475.wav"
+            ]
+        }
+    },
+    {
+        "id": 14077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003178.wav"
+            ]
+        }
+    },
+    {
+        "id": 14078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003846.wav"
+            ]
+        }
+    },
+    {
+        "id": 14081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003540.wav"
+            ]
+        }
+    },
+    {
+        "id": 14083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005414.wav"
+            ]
+        }
+    },
+    {
+        "id": 14084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003352.wav"
+            ]
+        }
+    },
+    {
+        "id": 14091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007244.wav"
+            ]
+        }
+    },
+    {
+        "id": 14101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004550.wav"
+            ]
+        }
+    },
+    {
+        "id": 14102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006265.wav"
+            ]
+        }
+    },
+    {
+        "id": 14109,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000548.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000548.wav"
+            ]
+        }
+    },
+    {
+        "id": 14110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008610.wav"
+            ]
+        }
+    },
+    {
+        "id": 14118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000809.wav"
+            ]
+        }
+    },
+    {
+        "id": 14131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007981.wav"
+            ]
+        }
+    },
+    {
+        "id": 14136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002051.wav"
+            ]
+        }
+    },
+    {
+        "id": 14137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002178.wav"
+            ]
+        }
+    },
+    {
+        "id": 14144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000843.wav"
+            ]
+        }
+    },
+    {
+        "id": 14147,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001656.wav"
+            ]
+        }
+    },
+    {
+        "id": 14149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006101.wav"
+            ]
+        }
+    },
+    {
+        "id": 14152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass on the left louder than the drum on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006934.wav"
+            ]
+        }
+    },
+    {
+        "id": 14154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005185.wav"
+            ]
+        }
+    },
+    {
+        "id": 14157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007230.wav"
+            ]
+        }
+    },
+    {
+        "id": 14169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006452.wav"
+            ]
+        }
+    },
+    {
+        "id": 14175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001334.wav"
+            ]
+        }
+    },
+    {
+        "id": 14183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003397.wav"
+            ]
+        }
+    },
+    {
+        "id": 14185,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005051.wav"
+            ]
+        }
+    },
+    {
+        "id": 14186,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006470.wav"
+            ]
+        }
+    },
+    {
+        "id": 14190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003512.wav"
+            ]
+        }
+    },
+    {
+        "id": 14198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005428.wav"
+            ]
+        }
+    },
+    {
+        "id": 14199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005392.wav"
+            ]
+        }
+    },
+    {
+        "id": 14205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005809.wav"
+            ]
+        }
+    },
+    {
+        "id": 14210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006052.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006052.wav"
+            ]
+        }
+    },
+    {
+        "id": 14215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the tuba on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007998.wav"
+            ]
+        }
+    },
+    {
+        "id": 14216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005840.wav"
+            ]
+        }
+    },
+    {
+        "id": 14217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000719.wav"
+            ]
+        }
+    },
+    {
+        "id": 14220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004266.wav"
+            ]
+        }
+    },
+    {
+        "id": 14233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005275.wav"
+            ]
+        }
+    },
+    {
+        "id": 14240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003879.wav"
+            ]
+        }
+    },
+    {
+        "id": 14250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001440.wav"
+            ]
+        }
+    },
+    {
+        "id": 14251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001800.wav"
+            ]
+        }
+    },
+    {
+        "id": 14254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007180.wav"
+            ]
+        }
+    },
+    {
+        "id": 14255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007522.wav"
+            ]
+        }
+    },
+    {
+        "id": 14259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004169.wav"
+            ]
+        }
+    },
+    {
+        "id": 14261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004920.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004920.wav"
+            ]
+        }
+    },
+    {
+        "id": 14267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006526.wav"
+            ]
+        }
+    },
+    {
+        "id": 14273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003800.wav"
+            ]
+        }
+    },
+    {
+        "id": 14277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the middle louder than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008044.wav"
+            ]
+        }
+    },
+    {
+        "id": 14281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005425.wav"
+            ]
+        }
+    },
+    {
+        "id": 14285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the middle louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008200.wav"
+            ]
+        }
+    },
+    {
+        "id": 14307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006579.wav"
+            ]
+        }
+    },
+    {
+        "id": 14308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003810.wav"
+            ]
+        }
+    },
+    {
+        "id": 14313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003329.wav"
+            ]
+        }
+    },
+    {
+        "id": 14319,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005468.wav"
+            ]
+        }
+    },
+    {
+        "id": 14320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000286.wav"
+            ]
+        }
+    },
+    {
+        "id": 14327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001264.wav"
+            ]
+        }
+    },
+    {
+        "id": 14337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004089.wav"
+            ]
+        }
+    },
+    {
+        "id": 14338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003619.wav"
+            ]
+        }
+    },
+    {
+        "id": 14339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003139.wav"
+            ]
+        }
+    },
+    {
+        "id": 14341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005432.wav"
+            ]
+        }
+    },
+    {
+        "id": 14346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001183.wav"
+            ]
+        }
+    },
+    {
+        "id": 14352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007420.wav"
+            ]
+        }
+    },
+    {
+        "id": 14357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008491.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008491.wav"
+            ]
+        }
+    },
+    {
+        "id": 14362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004575.wav"
+            ]
+        }
+    },
+    {
+        "id": 14366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007042.wav"
+            ]
+        }
+    },
+    {
+        "id": 14376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003247.wav"
+            ]
+        }
+    },
+    {
+        "id": 14381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the banjo on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001898.wav"
+            ]
+        }
+    },
+    {
+        "id": 14385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001207.wav"
+            ]
+        }
+    },
+    {
+        "id": 14399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006563.wav"
+            ]
+        }
+    },
+    {
+        "id": 14400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003042.wav"
+            ]
+        }
+    },
+    {
+        "id": 14405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005445.wav"
+            ]
+        }
+    },
+    {
+        "id": 14414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002104.wav"
+            ]
+        }
+    },
+    {
+        "id": 14418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002351.wav"
+            ]
+        }
+    },
+    {
+        "id": 14421,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004728.wav"
+            ]
+        }
+    },
+    {
+        "id": 14423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005808.wav"
+            ]
+        }
+    },
+    {
+        "id": 14424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003367.wav"
+            ]
+        }
+    },
+    {
+        "id": 14426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000130.wav"
+            ]
+        }
+    },
+    {
+        "id": 14438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002162.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002162.wav"
+            ]
+        }
+    },
+    {
+        "id": 14453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001077.wav"
+            ]
+        }
+    },
+    {
+        "id": 14456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007683.wav"
+            ]
+        }
+    },
+    {
+        "id": 14485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002930.wav"
+            ]
+        }
+    },
+    {
+        "id": 14493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006596.wav"
+            ]
+        }
+    },
+    {
+        "id": 14496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002201.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002201.wav"
+            ]
+        }
+    },
+    {
+        "id": 14508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002232.wav"
+            ]
+        }
+    },
+    {
+        "id": 14510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001532.wav"
+            ]
+        }
+    },
+    {
+        "id": 14512,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006060.wav"
+            ]
+        }
+    },
+    {
+        "id": 14515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004944.wav"
+            ]
+        }
+    },
+    {
+        "id": 14522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007813.wav"
+            ]
+        }
+    },
+    {
+        "id": 14529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005178.wav"
+            ]
+        }
+    },
+    {
+        "id": 14530,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002731.wav"
+            ]
+        }
+    },
+    {
+        "id": 14532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003901.wav"
+            ]
+        }
+    },
+    {
+        "id": 14533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005955.wav"
+            ]
+        }
+    },
+    {
+        "id": 14559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the middle louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008454.wav"
+            ]
+        }
+    },
+    {
+        "id": 14561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000212.wav"
+            ]
+        }
+    },
+    {
+        "id": 14562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003451.wav"
+            ]
+        }
+    },
+    {
+        "id": 14567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007059.wav"
+            ]
+        }
+    },
+    {
+        "id": 14572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002460.wav"
+            ]
+        }
+    },
+    {
+        "id": 14575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008449.wav"
+            ]
+        }
+    },
+    {
+        "id": 14578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006819.wav"
+            ]
+        }
+    },
+    {
+        "id": 14580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many suona are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006076.wav"
+            ]
+        }
+    },
+    {
+        "id": 14582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005061.wav"
+            ]
+        }
+    },
+    {
+        "id": 14585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003734.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003734.wav"
+            ]
+        }
+    },
+    {
+        "id": 14586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005077.wav"
+            ]
+        }
+    },
+    {
+        "id": 14587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000311.wav"
+            ]
+        }
+    },
+    {
+        "id": 14600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003651.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003651.wav"
+            ]
+        }
+    },
+    {
+        "id": 14601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008341.wav"
+            ]
+        }
+    },
+    {
+        "id": 14603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007966.wav"
+            ]
+        }
+    },
+    {
+        "id": 14612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003958.wav"
+            ]
+        }
+    },
+    {
+        "id": 14613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003049.wav"
+            ]
+        }
+    },
+    {
+        "id": 14616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004049.wav"
+            ]
+        }
+    },
+    {
+        "id": 14623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006499.wav"
+            ]
+        }
+    },
+    {
+        "id": 14625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003087.wav"
+            ]
+        }
+    },
+    {
+        "id": 14629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008558.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008558.wav"
+            ]
+        }
+    },
+    {
+        "id": 14631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000488.wav"
+            ]
+        }
+    },
+    {
+        "id": 14644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007343.wav"
+            ]
+        }
+    },
+    {
+        "id": 14654,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000441.wav"
+            ]
+        }
+    },
+    {
+        "id": 14655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004145.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004145.wav"
+            ]
+        }
+    },
+    {
+        "id": 14665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000092.wav"
+            ]
+        }
+    },
+    {
+        "id": 14681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001639.wav"
+            ]
+        }
+    },
+    {
+        "id": 14686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005422.wav"
+            ]
+        }
+    },
+    {
+        "id": 14695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000630.wav"
+            ]
+        }
+    },
+    {
+        "id": 14696,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007293.wav"
+            ]
+        }
+    },
+    {
+        "id": 14704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000481.wav"
+            ]
+        }
+    },
+    {
+        "id": 14707,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001614.wav"
+            ]
+        }
+    },
+    {
+        "id": 14708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003599.wav"
+            ]
+        }
+    },
+    {
+        "id": 14712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008022.wav"
+            ]
+        }
+    },
+    {
+        "id": 14713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008887.wav"
+            ]
+        }
+    },
+    {
+        "id": 14727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007038.wav"
+            ]
+        }
+    },
+    {
+        "id": 14733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006830.wav"
+            ]
+        }
+    },
+    {
+        "id": 14735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006552.wav"
+            ]
+        }
+    },
+    {
+        "id": 14737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004625.wav"
+            ]
+        }
+    },
+    {
+        "id": 14739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007099.wav"
+            ]
+        }
+    },
+    {
+        "id": 14751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006324.wav"
+            ]
+        }
+    },
+    {
+        "id": 14752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003841.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003841.wav"
+            ]
+        }
+    },
+    {
+        "id": 14764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008557.wav"
+            ]
+        }
+    },
+    {
+        "id": 14765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007388.wav"
+            ]
+        }
+    },
+    {
+        "id": 14767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000464.wav"
+            ]
+        }
+    },
+    {
+        "id": 14775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003631.wav"
+            ]
+        }
+    },
+    {
+        "id": 14777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002165.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002165.wav"
+            ]
+        }
+    },
+    {
+        "id": 14781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007142.wav"
+            ]
+        }
+    },
+    {
+        "id": 14784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002893.wav"
+            ]
+        }
+    },
+    {
+        "id": 14787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006929.wav"
+            ]
+        }
+    },
+    {
+        "id": 14790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008335.wav"
+            ]
+        }
+    },
+    {
+        "id": 14796,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008273.wav"
+            ]
+        }
+    },
+    {
+        "id": 14801,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003197.wav"
+            ]
+        }
+    },
+    {
+        "id": 14802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007450.wav"
+            ]
+        }
+    },
+    {
+        "id": 14803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008823.wav"
+            ]
+        }
+    },
+    {
+        "id": 14807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005343.wav"
+            ]
+        }
+    },
+    {
+        "id": 14817,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003805.wav"
+            ]
+        }
+    },
+    {
+        "id": 14821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004589.wav"
+            ]
+        }
+    },
+    {
+        "id": 14824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005922.wav"
+            ]
+        }
+    },
+    {
+        "id": 14826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006707.wav"
+            ]
+        }
+    },
+    {
+        "id": 14828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002284.wav"
+            ]
+        }
+    },
+    {
+        "id": 14832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003984.wav"
+            ]
+        }
+    },
+    {
+        "id": 14839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006749.wav"
+            ]
+        }
+    },
+    {
+        "id": 14848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006903.wav"
+            ]
+        }
+    },
+    {
+        "id": 14867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001168.wav"
+            ]
+        }
+    },
+    {
+        "id": 14875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004700.wav"
+            ]
+        }
+    },
+    {
+        "id": 14881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002079.wav"
+            ]
+        }
+    },
+    {
+        "id": 14883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008098.wav"
+            ]
+        }
+    },
+    {
+        "id": 14886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006325.wav"
+            ]
+        }
+    },
+    {
+        "id": 14892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008624.wav"
+            ]
+        }
+    },
+    {
+        "id": 14900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008576.wav"
+            ]
+        }
+    },
+    {
+        "id": 14908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008855.wav"
+            ]
+        }
+    },
+    {
+        "id": 14919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001422.wav"
+            ]
+        }
+    },
+    {
+        "id": 14929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000618.wav"
+            ]
+        }
+    },
+    {
+        "id": 14938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001938.wav"
+            ]
+        }
+    },
+    {
+        "id": 14940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000944.wav"
+            ]
+        }
+    },
+    {
+        "id": 14948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004035.wav"
+            ]
+        }
+    },
+    {
+        "id": 14949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004158.wav"
+            ]
+        }
+    },
+    {
+        "id": 14967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005960.wav"
+            ]
+        }
+    },
+    {
+        "id": 14970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004266.wav"
+            ]
+        }
+    },
+    {
+        "id": 14971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006843.wav"
+            ]
+        }
+    },
+    {
+        "id": 14981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001620.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001620.wav"
+            ]
+        }
+    },
+    {
+        "id": 14982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000483.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000483.wav"
+            ]
+        }
+    },
+    {
+        "id": 14984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007833.wav"
+            ]
+        }
+    },
+    {
+        "id": 14985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007499.wav"
+            ]
+        }
+    },
+    {
+        "id": 14987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000130.wav"
+            ]
+        }
+    },
+    {
+        "id": 14994,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007497.wav"
+            ]
+        }
+    },
+    {
+        "id": 14999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008065.wav"
+            ]
+        }
+    },
+    {
+        "id": 15008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007869.wav"
+            ]
+        }
+    },
+    {
+        "id": 15011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004346.wav"
+            ]
+        }
+    },
+    {
+        "id": 15016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008844.wav"
+            ]
+        }
+    },
+    {
+        "id": 15021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000385.wav"
+            ]
+        }
+    },
+    {
+        "id": 15027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008088.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008088.wav"
+            ]
+        }
+    },
+    {
+        "id": 15039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007355.wav"
+            ]
+        }
+    },
+    {
+        "id": 15041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003308.wav"
+            ]
+        }
+    },
+    {
+        "id": 15042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001265.wav"
+            ]
+        }
+    },
+    {
+        "id": 15054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002898.wav"
+            ]
+        }
+    },
+    {
+        "id": 15055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004158.wav"
+            ]
+        }
+    },
+    {
+        "id": 15058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000470.wav"
+            ]
+        }
+    },
+    {
+        "id": 15064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa playing longer than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002941.wav"
+            ]
+        }
+    },
+    {
+        "id": 15066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002417.wav"
+            ]
+        }
+    },
+    {
+        "id": 15067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002398.wav"
+            ]
+        }
+    },
+    {
+        "id": 15075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006089.wav"
+            ]
+        }
+    },
+    {
+        "id": 15079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002066.wav"
+            ]
+        }
+    },
+    {
+        "id": 15083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003979.wav"
+            ]
+        }
+    },
+    {
+        "id": 15089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007336.wav"
+            ]
+        }
+    },
+    {
+        "id": 15092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004798.wav"
+            ]
+        }
+    },
+    {
+        "id": 15105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005418.wav"
+            ]
+        }
+    },
+    {
+        "id": 15116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006522.wav"
+            ]
+        }
+    },
+    {
+        "id": 15124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005967.wav"
+            ]
+        }
+    },
+    {
+        "id": 15133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004457.wav"
+            ]
+        }
+    },
+    {
+        "id": 15139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004311.wav"
+            ]
+        }
+    },
+    {
+        "id": 15145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002495.wav"
+            ]
+        }
+    },
+    {
+        "id": 15152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008558.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008558.wav"
+            ]
+        }
+    },
+    {
+        "id": 15156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007156.wav"
+            ]
+        }
+    },
+    {
+        "id": 15168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005398.wav"
+            ]
+        }
+    },
+    {
+        "id": 15171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004456.wav"
+            ]
+        }
+    },
+    {
+        "id": 15175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002722.wav"
+            ]
+        }
+    },
+    {
+        "id": 15176,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000266.wav"
+            ]
+        }
+    },
+    {
+        "id": 15179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008354.wav"
+            ]
+        }
+    },
+    {
+        "id": 15185,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007953.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007953.wav"
+            ]
+        }
+    },
+    {
+        "id": 15187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007805.wav"
+            ]
+        }
+    },
+    {
+        "id": 15191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001300.wav"
+            ]
+        }
+    },
+    {
+        "id": 15194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002262.wav"
+            ]
+        }
+    },
+    {
+        "id": 15202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003073.wav"
+            ]
+        }
+    },
+    {
+        "id": 15212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008780.wav"
+            ]
+        }
+    },
+    {
+        "id": 15214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008091.wav"
+            ]
+        }
+    },
+    {
+        "id": 15218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002301.wav"
+            ]
+        }
+    },
+    {
+        "id": 15221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006758.wav"
+            ]
+        }
+    },
+    {
+        "id": 15233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006710.wav"
+            ]
+        }
+    },
+    {
+        "id": 15236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005775.wav"
+            ]
+        }
+    },
+    {
+        "id": 15238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006843.wav"
+            ]
+        }
+    },
+    {
+        "id": 15239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005974.wav"
+            ]
+        }
+    },
+    {
+        "id": 15242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002984.wav"
+            ]
+        }
+    },
+    {
+        "id": 15249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008512.wav"
+            ]
+        }
+    },
+    {
+        "id": 15252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004409.wav"
+            ]
+        }
+    },
+    {
+        "id": 15255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004417.wav"
+            ]
+        }
+    },
+    {
+        "id": 15256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001600.wav"
+            ]
+        }
+    },
+    {
+        "id": 15264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008925.wav"
+            ]
+        }
+    },
+    {
+        "id": 15267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006872.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006872.wav"
+            ]
+        }
+    },
+    {
+        "id": 15268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002604.wav"
+            ]
+        }
+    },
+    {
+        "id": 15273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007999.wav"
+            ]
+        }
+    },
+    {
+        "id": 15279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000055.wav"
+            ]
+        }
+    },
+    {
+        "id": 15291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006315.wav"
+            ]
+        }
+    },
+    {
+        "id": 15296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004306.wav"
+            ]
+        }
+    },
+    {
+        "id": 15302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006138.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006138.wav"
+            ]
+        }
+    },
+    {
+        "id": 15314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004164.wav"
+            ]
+        }
+    },
+    {
+        "id": 15320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002598.wav"
+            ]
+        }
+    },
+    {
+        "id": 15331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006699.wav"
+            ]
+        }
+    },
+    {
+        "id": 15339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000875.wav"
+            ]
+        }
+    },
+    {
+        "id": 15348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003752.wav"
+            ]
+        }
+    },
+    {
+        "id": 15349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right more rhythmic than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004466.wav"
+            ]
+        }
+    },
+    {
+        "id": 15362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001382.wav"
+            ]
+        }
+    },
+    {
+        "id": 15367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003615.wav"
+            ]
+        }
+    },
+    {
+        "id": 15371,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008399.wav"
+            ]
+        }
+    },
+    {
+        "id": 15383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006967.wav"
+            ]
+        }
+    },
+    {
+        "id": 15389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003237.wav"
+            ]
+        }
+    },
+    {
+        "id": 15399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004109.wav"
+            ]
+        }
+    },
+    {
+        "id": 15404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001107.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001107.wav"
+            ]
+        }
+    },
+    {
+        "id": 15406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007401.wav"
+            ]
+        }
+    },
+    {
+        "id": 15408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004839.wav"
+            ]
+        }
+    },
+    {
+        "id": 15410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004247.wav"
+            ]
+        }
+    },
+    {
+        "id": 15416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004587.wav"
+            ]
+        }
+    },
+    {
+        "id": 15422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001903.wav"
+            ]
+        }
+    },
+    {
+        "id": 15423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002856.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002856.wav"
+            ]
+        }
+    },
+    {
+        "id": 15424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006034.wav"
+            ]
+        }
+    },
+    {
+        "id": 15427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004333.wav"
+            ]
+        }
+    },
+    {
+        "id": 15434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003545.wav"
+            ]
+        }
+    },
+    {
+        "id": 15439,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001295.wav"
+            ]
+        }
+    },
+    {
+        "id": 15446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006371.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006371.wav"
+            ]
+        }
+    },
+    {
+        "id": 15457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007855.wav"
+            ]
+        }
+    },
+    {
+        "id": 15470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004648.wav"
+            ]
+        }
+    },
+    {
+        "id": 15475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004992.wav"
+            ]
+        }
+    },
+    {
+        "id": 15479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008495.wav"
+            ]
+        }
+    },
+    {
+        "id": 15481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007941.wav"
+            ]
+        }
+    },
+    {
+        "id": 15484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001653.wav"
+            ]
+        }
+    },
+    {
+        "id": 15490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002385.wav"
+            ]
+        }
+    },
+    {
+        "id": 15499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006434.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006434.wav"
+            ]
+        }
+    },
+    {
+        "id": 15507,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008591.wav"
+            ]
+        }
+    },
+    {
+        "id": 15512,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002819.wav"
+            ]
+        }
+    },
+    {
+        "id": 15514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the right louder than the pipa on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001732.wav"
+            ]
+        }
+    },
+    {
+        "id": 15515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003821.wav"
+            ]
+        }
+    },
+    {
+        "id": 15516,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008139.wav"
+            ]
+        }
+    },
+    {
+        "id": 15520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000330.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000330.wav"
+            ]
+        }
+    },
+    {
+        "id": 15538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007577.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007577.wav"
+            ]
+        }
+    },
+    {
+        "id": 15540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006920.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006920.wav"
+            ]
+        }
+    },
+    {
+        "id": 15543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004560.wav"
+            ]
+        }
+    },
+    {
+        "id": 15551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005799.wav"
+            ]
+        }
+    },
+    {
+        "id": 15552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006149.wav"
+            ]
+        }
+    },
+    {
+        "id": 15555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004095.wav"
+            ]
+        }
+    },
+    {
+        "id": 15560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002526.wav"
+            ]
+        }
+    },
+    {
+        "id": 15565,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007767.wav"
+            ]
+        }
+    },
+    {
+        "id": 15566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000392.wav"
+            ]
+        }
+    },
+    {
+        "id": 15570,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004870.wav"
+            ]
+        }
+    },
+    {
+        "id": 15574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005888.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005888.wav"
+            ]
+        }
+    },
+    {
+        "id": 15582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006035.wav"
+            ]
+        }
+    },
+    {
+        "id": 15595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004225.wav"
+            ]
+        }
+    },
+    {
+        "id": 15597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the right louder than the clarinet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008957.wav"
+            ]
+        }
+    },
+    {
+        "id": 15606,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002472.wav"
+            ]
+        }
+    },
+    {
+        "id": 15616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007765.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007765.wav"
+            ]
+        }
+    },
+    {
+        "id": 15620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000151.wav"
+            ]
+        }
+    },
+    {
+        "id": 15623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007231.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007231.wav"
+            ]
+        }
+    },
+    {
+        "id": 15629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004688.wav"
+            ]
+        }
+    },
+    {
+        "id": 15632,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005081.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005081.wav"
+            ]
+        }
+    },
+    {
+        "id": 15647,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006066.wav"
+            ]
+        }
+    },
+    {
+        "id": 15649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004715.wav"
+            ]
+        }
+    },
+    {
+        "id": 15650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006633.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006633.wav"
+            ]
+        }
+    },
+    {
+        "id": 15655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007619.wav"
+            ]
+        }
+    },
+    {
+        "id": 15666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002602.wav"
+            ]
+        }
+    },
+    {
+        "id": 15668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008616.wav"
+            ]
+        }
+    },
+    {
+        "id": 15670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004292.wav"
+            ]
+        }
+    },
+    {
+        "id": 15684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001639.wav"
+            ]
+        }
+    },
+    {
+        "id": 15685,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001769.wav"
+            ]
+        }
+    },
+    {
+        "id": 15704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002726.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002726.wav"
+            ]
+        }
+    },
+    {
+        "id": 15717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001284.wav"
+            ]
+        }
+    },
+    {
+        "id": 15718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bagpipe makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003143.wav"
+            ]
+        }
+    },
+    {
+        "id": 15719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000820.wav"
+            ]
+        }
+    },
+    {
+        "id": 15725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005979.wav"
+            ]
+        }
+    },
+    {
+        "id": 15733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008584.wav"
+            ]
+        }
+    },
+    {
+        "id": 15734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001892.wav"
+            ]
+        }
+    },
+    {
+        "id": 15738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002973.wav"
+            ]
+        }
+    },
+    {
+        "id": 15746,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002615.wav"
+            ]
+        }
+    },
+    {
+        "id": 15747,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000237.wav"
+            ]
+        }
+    },
+    {
+        "id": 15750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008753.wav"
+            ]
+        }
+    },
+    {
+        "id": 15752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003171.wav"
+            ]
+        }
+    },
+    {
+        "id": 15756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005344.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005344.wav"
+            ]
+        }
+    },
+    {
+        "id": 15776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002326.wav"
+            ]
+        }
+    },
+    {
+        "id": 15780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005025.wav"
+            ]
+        }
+    },
+    {
+        "id": 15781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004217.wav"
+            ]
+        }
+    },
+    {
+        "id": 15784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005614.wav"
+            ]
+        }
+    },
+    {
+        "id": 15788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005924.wav"
+            ]
+        }
+    },
+    {
+        "id": 15789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the right more rhythmic than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006082.wav"
+            ]
+        }
+    },
+    {
+        "id": 15792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006105.wav"
+            ]
+        }
+    },
+    {
+        "id": 15794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001984.wav"
+            ]
+        }
+    },
+    {
+        "id": 15812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002805.wav"
+            ]
+        }
+    },
+    {
+        "id": 15816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000797.wav"
+            ]
+        }
+    },
+    {
+        "id": 15824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004364.wav"
+            ]
+        }
+    },
+    {
+        "id": 15836,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003508.wav"
+            ]
+        }
+    },
+    {
+        "id": 15837,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004731.wav"
+            ]
+        }
+    },
+    {
+        "id": 15840,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005418.wav"
+            ]
+        }
+    },
+    {
+        "id": 15855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004233.wav"
+            ]
+        }
+    },
+    {
+        "id": 15863,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006864.wav"
+            ]
+        }
+    },
+    {
+        "id": 15878,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005115.wav"
+            ]
+        }
+    },
+    {
+        "id": 15879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002743.wav"
+            ]
+        }
+    },
+    {
+        "id": 15886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008809.wav"
+            ]
+        }
+    },
+    {
+        "id": 15889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005310.wav"
+            ]
+        }
+    },
+    {
+        "id": 15891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007395.wav"
+            ]
+        }
+    },
+    {
+        "id": 15898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008783.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008783.wav"
+            ]
+        }
+    },
+    {
+        "id": 15901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007344.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007344.wav"
+            ]
+        }
+    },
+    {
+        "id": 15907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002971.wav"
+            ]
+        }
+    },
+    {
+        "id": 15908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003160.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003160.wav"
+            ]
+        }
+    },
+    {
+        "id": 15909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008721.wav"
+            ]
+        }
+    },
+    {
+        "id": 15912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008362.wav"
+            ]
+        }
+    },
+    {
+        "id": 15921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005694.wav"
+            ]
+        }
+    },
+    {
+        "id": 15926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and xylophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008004.wav"
+            ]
+        }
+    },
+    {
+        "id": 15938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003246.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003246.wav"
+            ]
+        }
+    },
+    {
+        "id": 15939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005680.wav"
+            ]
+        }
+    },
+    {
+        "id": 15947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001351.wav"
+            ]
+        }
+    },
+    {
+        "id": 15953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004713.wav"
+            ]
+        }
+    },
+    {
+        "id": 15954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005306.wav"
+            ]
+        }
+    },
+    {
+        "id": 15960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006463.wav"
+            ]
+        }
+    },
+    {
+        "id": 15963,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003241.wav"
+            ]
+        }
+    },
+    {
+        "id": 15967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007190.wav"
+            ]
+        }
+    },
+    {
+        "id": 15972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008798.wav"
+            ]
+        }
+    },
+    {
+        "id": 15974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002157.wav"
+            ]
+        }
+    },
+    {
+        "id": 15987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005172.wav"
+            ]
+        }
+    },
+    {
+        "id": 15988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000913.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000913.wav"
+            ]
+        }
+    },
+    {
+        "id": 15995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001415.wav"
+            ]
+        }
+    },
+    {
+        "id": 15999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004597.wav"
+            ]
+        }
+    },
+    {
+        "id": 16002,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007979.wav"
+            ]
+        }
+    },
+    {
+        "id": 16004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007285.wav"
+            ]
+        }
+    },
+    {
+        "id": 16013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007003.wav"
+            ]
+        }
+    },
+    {
+        "id": 16018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004756.wav"
+            ]
+        }
+    },
+    {
+        "id": 16023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004672.wav"
+            ]
+        }
+    },
+    {
+        "id": 16024,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006464.wav"
+            ]
+        }
+    },
+    {
+        "id": 16028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006554.wav"
+            ]
+        }
+    },
+    {
+        "id": 16033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006749.wav"
+            ]
+        }
+    },
+    {
+        "id": 16045,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008200.wav"
+            ]
+        }
+    },
+    {
+        "id": 16047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000703.wav"
+            ]
+        }
+    },
+    {
+        "id": 16050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004232.wav"
+            ]
+        }
+    },
+    {
+        "id": 16052,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000153.wav"
+            ]
+        }
+    },
+    {
+        "id": 16054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002591.wav"
+            ]
+        }
+    },
+    {
+        "id": 16063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002854.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002854.wav"
+            ]
+        }
+    },
+    {
+        "id": 16066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005616.wav"
+            ]
+        }
+    },
+    {
+        "id": 16070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008485.wav"
+            ]
+        }
+    },
+    {
+        "id": 16072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003087.wav"
+            ]
+        }
+    },
+    {
+        "id": 16075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007728.wav"
+            ]
+        }
+    },
+    {
+        "id": 16080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003071.wav"
+            ]
+        }
+    },
+    {
+        "id": 16088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005450.wav"
+            ]
+        }
+    },
+    {
+        "id": 16097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005321.wav"
+            ]
+        }
+    },
+    {
+        "id": 16099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000652.wav"
+            ]
+        }
+    },
+    {
+        "id": 16101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005576.wav"
+            ]
+        }
+    },
+    {
+        "id": 16125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006275.wav"
+            ]
+        }
+    },
+    {
+        "id": 16136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002955.wav"
+            ]
+        }
+    },
+    {
+        "id": 16137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005315.wav"
+            ]
+        }
+    },
+    {
+        "id": 16143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000310.wav"
+            ]
+        }
+    },
+    {
+        "id": 16145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003416.wav"
+            ]
+        }
+    },
+    {
+        "id": 16153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007252.wav"
+            ]
+        }
+    },
+    {
+        "id": 16157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000512.wav"
+            ]
+        }
+    },
+    {
+        "id": 16166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001545.wav"
+            ]
+        }
+    },
+    {
+        "id": 16168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006478.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006478.wav"
+            ]
+        }
+    },
+    {
+        "id": 16174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000135.wav"
+            ]
+        }
+    },
+    {
+        "id": 16178,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005974.wav"
+            ]
+        }
+    },
+    {
+        "id": 16179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007533.wav"
+            ]
+        }
+    },
+    {
+        "id": 16180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006691.wav"
+            ]
+        }
+    },
+    {
+        "id": 16183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000610.wav"
+            ]
+        }
+    },
+    {
+        "id": 16184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006622.wav"
+            ]
+        }
+    },
+    {
+        "id": 16196,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006484.wav"
+            ]
+        }
+    },
+    {
+        "id": 16212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003292.wav"
+            ]
+        }
+    },
+    {
+        "id": 16214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002288.wav"
+            ]
+        }
+    },
+    {
+        "id": 16217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002342.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002342.wav"
+            ]
+        }
+    },
+    {
+        "id": 16224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005423.wav"
+            ]
+        }
+    },
+    {
+        "id": 16225,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008916.wav"
+            ]
+        }
+    },
+    {
+        "id": 16233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008318.wav"
+            ]
+        }
+    },
+    {
+        "id": 16237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007392.wav"
+            ]
+        }
+    },
+    {
+        "id": 16239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003876.wav"
+            ]
+        }
+    },
+    {
+        "id": 16241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008979.wav"
+            ]
+        }
+    },
+    {
+        "id": 16259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004043.wav"
+            ]
+        }
+    },
+    {
+        "id": 16264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001887.wav"
+            ]
+        }
+    },
+    {
+        "id": 16270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007819.wav"
+            ]
+        }
+    },
+    {
+        "id": 16275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007607.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007607.wav"
+            ]
+        }
+    },
+    {
+        "id": 16283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005687.wav"
+            ]
+        }
+    },
+    {
+        "id": 16284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005355.wav"
+            ]
+        }
+    },
+    {
+        "id": 16287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003166.wav"
+            ]
+        }
+    },
+    {
+        "id": 16293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000699.wav"
+            ]
+        }
+    },
+    {
+        "id": 16300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008846.wav"
+            ]
+        }
+    },
+    {
+        "id": 16304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004325.wav"
+            ]
+        }
+    },
+    {
+        "id": 16311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008288.wav"
+            ]
+        }
+    },
+    {
+        "id": 16313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005567.wav"
+            ]
+        }
+    },
+    {
+        "id": 16321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004405.wav"
+            ]
+        }
+    },
+    {
+        "id": 16325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005354.wav"
+            ]
+        }
+    },
+    {
+        "id": 16330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003355.wav"
+            ]
+        }
+    },
+    {
+        "id": 16334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000981.wav"
+            ]
+        }
+    },
+    {
+        "id": 16339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004119.wav"
+            ]
+        }
+    },
+    {
+        "id": 16342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000689.wav"
+            ]
+        }
+    },
+    {
+        "id": 16351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003287.wav"
+            ]
+        }
+    },
+    {
+        "id": 16352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004230.wav"
+            ]
+        }
+    },
+    {
+        "id": 16364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001121.wav"
+            ]
+        }
+    },
+    {
+        "id": 16401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008031.wav"
+            ]
+        }
+    },
+    {
+        "id": 16406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004035.wav"
+            ]
+        }
+    },
+    {
+        "id": 16411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000589.wav"
+            ]
+        }
+    },
+    {
+        "id": 16415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000723.wav"
+            ]
+        }
+    },
+    {
+        "id": 16417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000827.wav"
+            ]
+        }
+    },
+    {
+        "id": 16419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007863.wav"
+            ]
+        }
+    },
+    {
+        "id": 16423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007301.wav"
+            ]
+        }
+    },
+    {
+        "id": 16425,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002830.wav"
+            ]
+        }
+    },
+    {
+        "id": 16428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003776.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003776.wav"
+            ]
+        }
+    },
+    {
+        "id": 16429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000343.wav"
+            ]
+        }
+    },
+    {
+        "id": 16432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004653.wav"
+            ]
+        }
+    },
+    {
+        "id": 16435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004176.wav"
+            ]
+        }
+    },
+    {
+        "id": 16447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005697.wav"
+            ]
+        }
+    },
+    {
+        "id": 16448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007501.wav"
+            ]
+        }
+    },
+    {
+        "id": 16458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002790.wav"
+            ]
+        }
+    },
+    {
+        "id": 16462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000388.wav"
+            ]
+        }
+    },
+    {
+        "id": 16475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007441.wav"
+            ]
+        }
+    },
+    {
+        "id": 16477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002337.wav"
+            ]
+        }
+    },
+    {
+        "id": 16482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007219.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007219.wav"
+            ]
+        }
+    },
+    {
+        "id": 16491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007001.wav"
+            ]
+        }
+    },
+    {
+        "id": 16492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003638.wav"
+            ]
+        }
+    },
+    {
+        "id": 16498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000493.wav"
+            ]
+        }
+    },
+    {
+        "id": 16513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002870.wav"
+            ]
+        }
+    },
+    {
+        "id": 16524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003336.wav"
+            ]
+        }
+    },
+    {
+        "id": 16526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000112.wav"
+            ]
+        }
+    },
+    {
+        "id": 16533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006518.wav"
+            ]
+        }
+    },
+    {
+        "id": 16537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008855.wav"
+            ]
+        }
+    },
+    {
+        "id": 16538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001918.wav"
+            ]
+        }
+    },
+    {
+        "id": 16539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000471.wav"
+            ]
+        }
+    },
+    {
+        "id": 16540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002285.wav"
+            ]
+        }
+    },
+    {
+        "id": 16541,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001672.wav"
+            ]
+        }
+    },
+    {
+        "id": 16550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007314.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007314.wav"
+            ]
+        }
+    },
+    {
+        "id": 16556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000172.wav"
+            ]
+        }
+    },
+    {
+        "id": 16557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002406.wav"
+            ]
+        }
+    },
+    {
+        "id": 16563,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001586.wav"
+            ]
+        }
+    },
+    {
+        "id": 16564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001244.wav"
+            ]
+        }
+    },
+    {
+        "id": 16574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001814.wav"
+            ]
+        }
+    },
+    {
+        "id": 16579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000045.wav"
+            ]
+        }
+    },
+    {
+        "id": 16583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006696.wav"
+            ]
+        }
+    },
+    {
+        "id": 16584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005601.wav"
+            ]
+        }
+    },
+    {
+        "id": 16585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005878.wav"
+            ]
+        }
+    },
+    {
+        "id": 16593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003489.wav"
+            ]
+        }
+    },
+    {
+        "id": 16609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000023.wav"
+            ]
+        }
+    },
+    {
+        "id": 16612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003024.wav"
+            ]
+        }
+    },
+    {
+        "id": 16614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008711.wav"
+            ]
+        }
+    },
+    {
+        "id": 16617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005929.wav"
+            ]
+        }
+    },
+    {
+        "id": 16619,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005913.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005913.wav"
+            ]
+        }
+    },
+    {
+        "id": 16623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008595.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008595.wav"
+            ]
+        }
+    },
+    {
+        "id": 16630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002151.wav"
+            ]
+        }
+    },
+    {
+        "id": 16631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003806.wav"
+            ]
+        }
+    },
+    {
+        "id": 16633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006767.wav"
+            ]
+        }
+    },
+    {
+        "id": 16634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002370.wav"
+            ]
+        }
+    },
+    {
+        "id": 16635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006788.wav"
+            ]
+        }
+    },
+    {
+        "id": 16641,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000050.wav"
+            ]
+        }
+    },
+    {
+        "id": 16645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003858.wav"
+            ]
+        }
+    },
+    {
+        "id": 16646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001700.wav"
+            ]
+        }
+    },
+    {
+        "id": 16647,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right more rhythmic than the electric_bass on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008012.wav"
+            ]
+        }
+    },
+    {
+        "id": 16653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008363.wav"
+            ]
+        }
+    },
+    {
+        "id": 16655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008247.wav"
+            ]
+        }
+    },
+    {
+        "id": 16672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000591.wav"
+            ]
+        }
+    },
+    {
+        "id": 16680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008289.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008289.wav"
+            ]
+        }
+    },
+    {
+        "id": 16687,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007302.wav"
+            ]
+        }
+    },
+    {
+        "id": 16708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000862.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000862.wav"
+            ]
+        }
+    },
+    {
+        "id": 16720,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005133.wav"
+            ]
+        }
+    },
+    {
+        "id": 16726,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001619.wav"
+            ]
+        }
+    },
+    {
+        "id": 16734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000402.wav"
+            ]
+        }
+    },
+    {
+        "id": 16736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003043.wav"
+            ]
+        }
+    },
+    {
+        "id": 16740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006673.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006673.wav"
+            ]
+        }
+    },
+    {
+        "id": 16749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002859.wav"
+            ]
+        }
+    },
+    {
+        "id": 16750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000109.wav"
+            ]
+        }
+    },
+    {
+        "id": 16759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000771.wav"
+            ]
+        }
+    },
+    {
+        "id": 16764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008449.wav"
+            ]
+        }
+    },
+    {
+        "id": 16765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006530.wav"
+            ]
+        }
+    },
+    {
+        "id": 16776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many suona are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000168.wav"
+            ]
+        }
+    },
+    {
+        "id": 16777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008426.wav"
+            ]
+        }
+    },
+    {
+        "id": 16780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004005.wav"
+            ]
+        }
+    },
+    {
+        "id": 16793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001930.wav"
+            ]
+        }
+    },
+    {
+        "id": 16806,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004077.wav"
+            ]
+        }
+    },
+    {
+        "id": 16807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006517.wav"
+            ]
+        }
+    },
+    {
+        "id": 16812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004956.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004956.wav"
+            ]
+        }
+    },
+    {
+        "id": 16816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000751.wav"
+            ]
+        }
+    },
+    {
+        "id": 16826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001415.wav"
+            ]
+        }
+    },
+    {
+        "id": 16828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005361.wav"
+            ]
+        }
+    },
+    {
+        "id": 16843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004497.wav"
+            ]
+        }
+    },
+    {
+        "id": 16844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005144.wav"
+            ]
+        }
+    },
+    {
+        "id": 16845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001255.wav"
+            ]
+        }
+    },
+    {
+        "id": 16848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008516.wav"
+            ]
+        }
+    },
+    {
+        "id": 16849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006554.wav"
+            ]
+        }
+    },
+    {
+        "id": 16854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006019.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006019.wav"
+            ]
+        }
+    },
+    {
+        "id": 16872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000740.wav"
+            ]
+        }
+    },
+    {
+        "id": 16874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004912.wav"
+            ]
+        }
+    },
+    {
+        "id": 16875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004606.wav"
+            ]
+        }
+    },
+    {
+        "id": 16897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005137.wav"
+            ]
+        }
+    },
+    {
+        "id": 16903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001274.wav"
+            ]
+        }
+    },
+    {
+        "id": 16906,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000865.wav"
+            ]
+        }
+    },
+    {
+        "id": 16909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007546.wav"
+            ]
+        }
+    },
+    {
+        "id": 16919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005802.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005802.wav"
+            ]
+        }
+    },
+    {
+        "id": 16926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004333.wav"
+            ]
+        }
+    },
+    {
+        "id": 16934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002646.wav"
+            ]
+        }
+    },
+    {
+        "id": 16947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008799.wav"
+            ]
+        }
+    },
+    {
+        "id": 16949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000008.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000008.wav"
+            ]
+        }
+    },
+    {
+        "id": 16951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000882.wav"
+            ]
+        }
+    },
+    {
+        "id": 16961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001444.wav"
+            ]
+        }
+    },
+    {
+        "id": 16964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007940.wav"
+            ]
+        }
+    },
+    {
+        "id": 16969,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007116.wav"
+            ]
+        }
+    },
+    {
+        "id": 16970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003297.wav"
+            ]
+        }
+    },
+    {
+        "id": 16971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000807.wav"
+            ]
+        }
+    },
+    {
+        "id": 16976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003325.wav"
+            ]
+        }
+    },
+    {
+        "id": 16977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002657.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002657.wav"
+            ]
+        }
+    },
+    {
+        "id": 16979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004048.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004048.wav"
+            ]
+        }
+    },
+    {
+        "id": 16982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006041.wav"
+            ]
+        }
+    },
+    {
+        "id": 17020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002686.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002686.wav"
+            ]
+        }
+    },
+    {
+        "id": 17022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002460.wav"
+            ]
+        }
+    },
+    {
+        "id": 17026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005173.wav"
+            ]
+        }
+    },
+    {
+        "id": 17032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007377.wav"
+            ]
+        }
+    },
+    {
+        "id": 17033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002923.wav"
+            ]
+        }
+    },
+    {
+        "id": 17035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007189.wav"
+            ]
+        }
+    },
+    {
+        "id": 17036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003336.wav"
+            ]
+        }
+    },
+    {
+        "id": 17041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007095.wav"
+            ]
+        }
+    },
+    {
+        "id": 17048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008490.wav"
+            ]
+        }
+    },
+    {
+        "id": 17058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002897.wav"
+            ]
+        }
+    },
+    {
+        "id": 17069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000581.wav"
+            ]
+        }
+    },
+    {
+        "id": 17073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001151.wav"
+            ]
+        }
+    },
+    {
+        "id": 17074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001867.wav"
+            ]
+        }
+    },
+    {
+        "id": 17077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000315.wav"
+            ]
+        }
+    },
+    {
+        "id": 17085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe on the left louder than the bagpipe on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006741.wav"
+            ]
+        }
+    },
+    {
+        "id": 17091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003762.wav"
+            ]
+        }
+    },
+    {
+        "id": 17099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002071.wav"
+            ]
+        }
+    },
+    {
+        "id": 17104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004610.wav"
+            ]
+        }
+    },
+    {
+        "id": 17108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005629.wav"
+            ]
+        }
+    },
+    {
+        "id": 17112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003374.wav"
+            ]
+        }
+    },
+    {
+        "id": 17118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002805.wav"
+            ]
+        }
+    },
+    {
+        "id": 17122,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006823.wav"
+            ]
+        }
+    },
+    {
+        "id": 17124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006452.wav"
+            ]
+        }
+    },
+    {
+        "id": 17129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005852.wav"
+            ]
+        }
+    },
+    {
+        "id": 17131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left louder than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003513.wav"
+            ]
+        }
+    },
+    {
+        "id": 17132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000735.wav"
+            ]
+        }
+    },
+    {
+        "id": 17137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001625.wav"
+            ]
+        }
+    },
+    {
+        "id": 17158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001135.wav"
+            ]
+        }
+    },
+    {
+        "id": 17173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000948.wav"
+            ]
+        }
+    },
+    {
+        "id": 17189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005436.wav"
+            ]
+        }
+    },
+    {
+        "id": 17194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002699.wav"
+            ]
+        }
+    },
+    {
+        "id": 17196,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005804.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005804.wav"
+            ]
+        }
+    },
+    {
+        "id": 17200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001476.wav"
+            ]
+        }
+    },
+    {
+        "id": 17201,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007591.wav"
+            ]
+        }
+    },
+    {
+        "id": 17202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006039.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006039.wav"
+            ]
+        }
+    },
+    {
+        "id": 17210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007520.wav"
+            ]
+        }
+    },
+    {
+        "id": 17215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003984.wav"
+            ]
+        }
+    },
+    {
+        "id": 17218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005454.wav"
+            ]
+        }
+    },
+    {
+        "id": 17224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001070.wav"
+            ]
+        }
+    },
+    {
+        "id": 17228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007914.wav"
+            ]
+        }
+    },
+    {
+        "id": 17231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005411.wav"
+            ]
+        }
+    },
+    {
+        "id": 17235,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000163.wav"
+            ]
+        }
+    },
+    {
+        "id": 17240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005557.wav"
+            ]
+        }
+    },
+    {
+        "id": 17244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008464.wav"
+            ]
+        }
+    },
+    {
+        "id": 17249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005439.wav"
+            ]
+        }
+    },
+    {
+        "id": 17254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004757.wav"
+            ]
+        }
+    },
+    {
+        "id": 17255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006931.wav"
+            ]
+        }
+    },
+    {
+        "id": 17257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008874.wav"
+            ]
+        }
+    },
+    {
+        "id": 17263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000337.wav"
+            ]
+        }
+    },
+    {
+        "id": 17264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000573.wav"
+            ]
+        }
+    },
+    {
+        "id": 17270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007061.wav"
+            ]
+        }
+    },
+    {
+        "id": 17274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002365.wav"
+            ]
+        }
+    },
+    {
+        "id": 17281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008029.wav"
+            ]
+        }
+    },
+    {
+        "id": 17284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002619.wav"
+            ]
+        }
+    },
+    {
+        "id": 17285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008074.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008074.wav"
+            ]
+        }
+    },
+    {
+        "id": 17288,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003970.wav"
+            ]
+        }
+    },
+    {
+        "id": 17290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001484.wav"
+            ]
+        }
+    },
+    {
+        "id": 17301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006522.wav"
+            ]
+        }
+    },
+    {
+        "id": 17307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006355.wav"
+            ]
+        }
+    },
+    {
+        "id": 17313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004812.wav"
+            ]
+        }
+    },
+    {
+        "id": 17317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000036.wav"
+            ]
+        }
+    },
+    {
+        "id": 17323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000582.wav"
+            ]
+        }
+    },
+    {
+        "id": 17332,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003643.wav"
+            ]
+        }
+    },
+    {
+        "id": 17333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007631.wav"
+            ]
+        }
+    },
+    {
+        "id": 17338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006716.wav"
+            ]
+        }
+    },
+    {
+        "id": 17339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002372.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002372.wav"
+            ]
+        }
+    },
+    {
+        "id": 17345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007136.wav"
+            ]
+        }
+    },
+    {
+        "id": 17347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006992.wav"
+            ]
+        }
+    },
+    {
+        "id": 17353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005929.wav"
+            ]
+        }
+    },
+    {
+        "id": 17363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004439.wav"
+            ]
+        }
+    },
+    {
+        "id": 17366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002629.wav"
+            ]
+        }
+    },
+    {
+        "id": 17384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003553.wav"
+            ]
+        }
+    },
+    {
+        "id": 17390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003550.wav"
+            ]
+        }
+    },
+    {
+        "id": 17395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there congas and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008032.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008032.wav"
+            ]
+        }
+    },
+    {
+        "id": 17398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002211.wav"
+            ]
+        }
+    },
+    {
+        "id": 17417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002018.wav"
+            ]
+        }
+    },
+    {
+        "id": 17421,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005703.wav"
+            ]
+        }
+    },
+    {
+        "id": 17422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005109.wav"
+            ]
+        }
+    },
+    {
+        "id": 17424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001766.wav"
+            ]
+        }
+    },
+    {
+        "id": 17430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007850.wav"
+            ]
+        }
+    },
+    {
+        "id": 17434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007485.wav"
+            ]
+        }
+    },
+    {
+        "id": 17438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000080.wav"
+            ]
+        }
+    },
+    {
+        "id": 17443,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002994.wav"
+            ]
+        }
+    },
+    {
+        "id": 17448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001767.wav"
+            ]
+        }
+    },
+    {
+        "id": 17451,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008655.wav"
+            ]
+        }
+    },
+    {
+        "id": 17454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004986.wav"
+            ]
+        }
+    },
+    {
+        "id": 17458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005646.wav"
+            ]
+        }
+    },
+    {
+        "id": 17461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006648.wav"
+            ]
+        }
+    },
+    {
+        "id": 17463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008199.wav"
+            ]
+        }
+    },
+    {
+        "id": 17472,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008802.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008802.wav"
+            ]
+        }
+    },
+    {
+        "id": 17475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002702.wav"
+            ]
+        }
+    },
+    {
+        "id": 17479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006439.wav"
+            ]
+        }
+    },
+    {
+        "id": 17493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002091.wav"
+            ]
+        }
+    },
+    {
+        "id": 17505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005622.wav"
+            ]
+        }
+    },
+    {
+        "id": 17512,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002240.wav"
+            ]
+        }
+    },
+    {
+        "id": 17513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000976.wav"
+            ]
+        }
+    },
+    {
+        "id": 17520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002565.wav"
+            ]
+        }
+    },
+    {
+        "id": 17521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000654.wav"
+            ]
+        }
+    },
+    {
+        "id": 17524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007681.wav"
+            ]
+        }
+    },
+    {
+        "id": 17526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001296.wav"
+            ]
+        }
+    },
+    {
+        "id": 17529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich banjo makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003393.wav"
+            ]
+        }
+    },
+    {
+        "id": 17532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004262.wav"
+            ]
+        }
+    },
+    {
+        "id": 17543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002780.wav"
+            ]
+        }
+    },
+    {
+        "id": 17552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007490.wav"
+            ]
+        }
+    },
+    {
+        "id": 17562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001007.wav"
+            ]
+        }
+    },
+    {
+        "id": 17566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000448.wav"
+            ]
+        }
+    },
+    {
+        "id": 17569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000145.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000145.wav"
+            ]
+        }
+    },
+    {
+        "id": 17572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002044.wav"
+            ]
+        }
+    },
+    {
+        "id": 17575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the middle louder than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006596.wav"
+            ]
+        }
+    },
+    {
+        "id": 17577,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005463.wav"
+            ]
+        }
+    },
+    {
+        "id": 17579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007831.wav"
+            ]
+        }
+    },
+    {
+        "id": 17580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003399.wav"
+            ]
+        }
+    },
+    {
+        "id": 17581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002709.wav"
+            ]
+        }
+    },
+    {
+        "id": 17586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004739.wav"
+            ]
+        }
+    },
+    {
+        "id": 17595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003713.wav"
+            ]
+        }
+    },
+    {
+        "id": 17607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004487.wav"
+            ]
+        }
+    },
+    {
+        "id": 17611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000438.wav"
+            ]
+        }
+    },
+    {
+        "id": 17618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005035.wav"
+            ]
+        }
+    },
+    {
+        "id": 17621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001616.wav"
+            ]
+        }
+    },
+    {
+        "id": 17624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007065.wav"
+            ]
+        }
+    },
+    {
+        "id": 17627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006608.wav"
+            ]
+        }
+    },
+    {
+        "id": 17635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003143.wav"
+            ]
+        }
+    },
+    {
+        "id": 17642,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006999.wav"
+            ]
+        }
+    },
+    {
+        "id": 17644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005665.wav"
+            ]
+        }
+    },
+    {
+        "id": 17650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004842.wav"
+            ]
+        }
+    },
+    {
+        "id": 17653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001760.wav"
+            ]
+        }
+    },
+    {
+        "id": 17655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001157.wav"
+            ]
+        }
+    },
+    {
+        "id": 17670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008253.wav"
+            ]
+        }
+    },
+    {
+        "id": 17678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006677.wav"
+            ]
+        }
+    },
+    {
+        "id": 17693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008300.wav"
+            ]
+        }
+    },
+    {
+        "id": 17699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006897.wav"
+            ]
+        }
+    },
+    {
+        "id": 17704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006566.wav"
+            ]
+        }
+    },
+    {
+        "id": 17709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004055.wav"
+            ]
+        }
+    },
+    {
+        "id": 17729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001077.wav"
+            ]
+        }
+    },
+    {
+        "id": 17733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000481.wav"
+            ]
+        }
+    },
+    {
+        "id": 17742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003248.wav"
+            ]
+        }
+    },
+    {
+        "id": 17743,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005747.wav"
+            ]
+        }
+    },
+    {
+        "id": 17744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005535.wav"
+            ]
+        }
+    },
+    {
+        "id": 17748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004832.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004832.wav"
+            ]
+        }
+    },
+    {
+        "id": 17751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004796.wav"
+            ]
+        }
+    },
+    {
+        "id": 17754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005935.wav"
+            ]
+        }
+    },
+    {
+        "id": 17759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006231.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006231.wav"
+            ]
+        }
+    },
+    {
+        "id": 17767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000816.wav"
+            ]
+        }
+    },
+    {
+        "id": 17782,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004503.wav"
+            ]
+        }
+    },
+    {
+        "id": 17784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006482.wav"
+            ]
+        }
+    },
+    {
+        "id": 17786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004183.wav"
+            ]
+        }
+    },
+    {
+        "id": 17787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007339.wav"
+            ]
+        }
+    },
+    {
+        "id": 17788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the saxophone on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007868.wav"
+            ]
+        }
+    },
+    {
+        "id": 17792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona on the middle louder than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007684.wav"
+            ]
+        }
+    },
+    {
+        "id": 17795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006713.wav"
+            ]
+        }
+    },
+    {
+        "id": 17798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008575.wav"
+            ]
+        }
+    },
+    {
+        "id": 17802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003215.wav"
+            ]
+        }
+    },
+    {
+        "id": 17807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005334.wav"
+            ]
+        }
+    },
+    {
+        "id": 17813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008188.wav"
+            ]
+        }
+    },
+    {
+        "id": 17816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004535.wav"
+            ]
+        }
+    },
+    {
+        "id": 17821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002261.wav"
+            ]
+        }
+    },
+    {
+        "id": 17832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003261.wav"
+            ]
+        }
+    },
+    {
+        "id": 17833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005353.wav"
+            ]
+        }
+    },
+    {
+        "id": 17841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005350.wav"
+            ]
+        }
+    },
+    {
+        "id": 17842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000289.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000289.wav"
+            ]
+        }
+    },
+    {
+        "id": 17843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003847.wav"
+            ]
+        }
+    },
+    {
+        "id": 17846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004150.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004150.wav"
+            ]
+        }
+    },
+    {
+        "id": 17850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004216.wav"
+            ]
+        }
+    },
+    {
+        "id": 17857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007970.wav"
+            ]
+        }
+    },
+    {
+        "id": 17864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002967.wav"
+            ]
+        }
+    },
+    {
+        "id": 17867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001181.wav"
+            ]
+        }
+    },
+    {
+        "id": 17868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008348.wav"
+            ]
+        }
+    },
+    {
+        "id": 17879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the accordion on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002686.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002686.wav"
+            ]
+        }
+    },
+    {
+        "id": 17880,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008729.wav"
+            ]
+        }
+    },
+    {
+        "id": 17885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001775.wav"
+            ]
+        }
+    },
+    {
+        "id": 17893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004390.wav"
+            ]
+        }
+    },
+    {
+        "id": 17907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003978.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003978.wav"
+            ]
+        }
+    },
+    {
+        "id": 17909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001333.wav"
+            ]
+        }
+    },
+    {
+        "id": 17911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007965.wav"
+            ]
+        }
+    },
+    {
+        "id": 17917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001007.wav"
+            ]
+        }
+    },
+    {
+        "id": 17922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007858.wav"
+            ]
+        }
+    },
+    {
+        "id": 17925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003983.wav"
+            ]
+        }
+    },
+    {
+        "id": 17928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004675.wav"
+            ]
+        }
+    },
+    {
+        "id": 17941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003085.wav"
+            ]
+        }
+    },
+    {
+        "id": 17943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004638.wav"
+            ]
+        }
+    },
+    {
+        "id": 17947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003099.wav"
+            ]
+        }
+    },
+    {
+        "id": 17949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002059.wav"
+            ]
+        }
+    },
+    {
+        "id": 17950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007549.wav"
+            ]
+        }
+    },
+    {
+        "id": 17951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007357.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007357.wav"
+            ]
+        }
+    },
+    {
+        "id": 17953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004741.wav"
+            ]
+        }
+    },
+    {
+        "id": 17967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007426.wav"
+            ]
+        }
+    },
+    {
+        "id": 17972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the middle louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004910.wav"
+            ]
+        }
+    },
+    {
+        "id": 17977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000844.wav"
+            ]
+        }
+    },
+    {
+        "id": 17978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000463.wav"
+            ]
+        }
+    },
+    {
+        "id": 17983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007865.wav"
+            ]
+        }
+    },
+    {
+        "id": 17993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006444.wav"
+            ]
+        }
+    },
+    {
+        "id": 17997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005023.wav"
+            ]
+        }
+    },
+    {
+        "id": 18013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005808.wav"
+            ]
+        }
+    },
+    {
+        "id": 18018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004455.wav"
+            ]
+        }
+    },
+    {
+        "id": 18024,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006189.wav"
+            ]
+        }
+    },
+    {
+        "id": 18028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000537.wav"
+            ]
+        }
+    },
+    {
+        "id": 18033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001076.wav"
+            ]
+        }
+    },
+    {
+        "id": 18034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005227.wav"
+            ]
+        }
+    },
+    {
+        "id": 18038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001212.wav"
+            ]
+        }
+    },
+    {
+        "id": 18042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000954.wav"
+            ]
+        }
+    },
+    {
+        "id": 18043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007157.wav"
+            ]
+        }
+    },
+    {
+        "id": 18047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007017.wav"
+            ]
+        }
+    },
+    {
+        "id": 18054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006555.wav"
+            ]
+        }
+    },
+    {
+        "id": 18059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000367.wav"
+            ]
+        }
+    },
+    {
+        "id": 18060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002160.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002160.wav"
+            ]
+        }
+    },
+    {
+        "id": 18061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007627.wav"
+            ]
+        }
+    },
+    {
+        "id": 18063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001552.wav"
+            ]
+        }
+    },
+    {
+        "id": 18070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001345.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001345.wav"
+            ]
+        }
+    },
+    {
+        "id": 18072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003885.wav"
+            ]
+        }
+    },
+    {
+        "id": 18076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004940.wav"
+            ]
+        }
+    },
+    {
+        "id": 18079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003489.wav"
+            ]
+        }
+    },
+    {
+        "id": 18080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006721.wav"
+            ]
+        }
+    },
+    {
+        "id": 18083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the clarinet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002143.wav"
+            ]
+        }
+    },
+    {
+        "id": 18091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003811.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003811.wav"
+            ]
+        }
+    },
+    {
+        "id": 18093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008051.wav"
+            ]
+        }
+    },
+    {
+        "id": 18096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006430.wav"
+            ]
+        }
+    },
+    {
+        "id": 18097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006790.wav"
+            ]
+        }
+    },
+    {
+        "id": 18101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001963.wav"
+            ]
+        }
+    },
+    {
+        "id": 18106,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006875.wav"
+            ]
+        }
+    },
+    {
+        "id": 18107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008709.wav"
+            ]
+        }
+    },
+    {
+        "id": 18110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008813.wav"
+            ]
+        }
+    },
+    {
+        "id": 18115,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003723.wav"
+            ]
+        }
+    },
+    {
+        "id": 18116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006640.wav"
+            ]
+        }
+    },
+    {
+        "id": 18128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007929.wav"
+            ]
+        }
+    },
+    {
+        "id": 18133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007600.wav"
+            ]
+        }
+    },
+    {
+        "id": 18141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008587.wav"
+            ]
+        }
+    },
+    {
+        "id": 18159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003319.wav"
+            ]
+        }
+    },
+    {
+        "id": 18161,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002842.wav"
+            ]
+        }
+    },
+    {
+        "id": 18167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005057.wav"
+            ]
+        }
+    },
+    {
+        "id": 18170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000292.wav"
+            ]
+        }
+    },
+    {
+        "id": 18185,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008791.wav"
+            ]
+        }
+    },
+    {
+        "id": 18187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007159.wav"
+            ]
+        }
+    },
+    {
+        "id": 18190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006744.wav"
+            ]
+        }
+    },
+    {
+        "id": 18198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006943.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006943.wav"
+            ]
+        }
+    },
+    {
+        "id": 18199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001481.wav"
+            ]
+        }
+    },
+    {
+        "id": 18205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002835.wav"
+            ]
+        }
+    },
+    {
+        "id": 18206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005604.wav"
+            ]
+        }
+    },
+    {
+        "id": 18207,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003678.wav"
+            ]
+        }
+    },
+    {
+        "id": 18216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000530.wav"
+            ]
+        }
+    },
+    {
+        "id": 18220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006078.wav"
+            ]
+        }
+    },
+    {
+        "id": 18224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007627.wav"
+            ]
+        }
+    },
+    {
+        "id": 18228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006233.wav"
+            ]
+        }
+    },
+    {
+        "id": 18234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008129.wav"
+            ]
+        }
+    },
+    {
+        "id": 18236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004925.wav"
+            ]
+        }
+    },
+    {
+        "id": 18251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001846.wav"
+            ]
+        }
+    },
+    {
+        "id": 18252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008705.wav"
+            ]
+        }
+    },
+    {
+        "id": 18253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001672.wav"
+            ]
+        }
+    },
+    {
+        "id": 18270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004122.wav"
+            ]
+        }
+    },
+    {
+        "id": 18273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008409.wav"
+            ]
+        }
+    },
+    {
+        "id": 18283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008444.wav"
+            ]
+        }
+    },
+    {
+        "id": 18291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008203.wav"
+            ]
+        }
+    },
+    {
+        "id": 18310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001075.wav"
+            ]
+        }
+    },
+    {
+        "id": 18311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003497.wav"
+            ]
+        }
+    },
+    {
+        "id": 18312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002439.wav"
+            ]
+        }
+    },
+    {
+        "id": 18316,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000068.wav"
+            ]
+        }
+    },
+    {
+        "id": 18323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004271.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004271.wav"
+            ]
+        }
+    },
+    {
+        "id": 18325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008478.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008478.wav"
+            ]
+        }
+    },
+    {
+        "id": 18327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000519.wav"
+            ]
+        }
+    },
+    {
+        "id": 18335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007163.wav"
+            ]
+        }
+    },
+    {
+        "id": 18341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007030.wav"
+            ]
+        }
+    },
+    {
+        "id": 18355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008923.wav"
+            ]
+        }
+    },
+    {
+        "id": 18356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004138.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004138.wav"
+            ]
+        }
+    },
+    {
+        "id": 18365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004949.wav"
+            ]
+        }
+    },
+    {
+        "id": 18366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001929.wav"
+            ]
+        }
+    },
+    {
+        "id": 18368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002261.wav"
+            ]
+        }
+    },
+    {
+        "id": 18373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006701.wav"
+            ]
+        }
+    },
+    {
+        "id": 18374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002499.wav"
+            ]
+        }
+    },
+    {
+        "id": 18378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000319.wav"
+            ]
+        }
+    },
+    {
+        "id": 18383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006025.wav"
+            ]
+        }
+    },
+    {
+        "id": 18384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006495.wav"
+            ]
+        }
+    },
+    {
+        "id": 18392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003947.wav"
+            ]
+        }
+    },
+    {
+        "id": 18397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008307.wav"
+            ]
+        }
+    },
+    {
+        "id": 18403,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004372.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004372.wav"
+            ]
+        }
+    },
+    {
+        "id": 18405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003009.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003009.wav"
+            ]
+        }
+    },
+    {
+        "id": 18409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004062.wav"
+            ]
+        }
+    },
+    {
+        "id": 18420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005789.wav"
+            ]
+        }
+    },
+    {
+        "id": 18422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001764.wav"
+            ]
+        }
+    },
+    {
+        "id": 18428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003590.wav"
+            ]
+        }
+    },
+    {
+        "id": 18433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004875.wav"
+            ]
+        }
+    },
+    {
+        "id": 18447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008571.wav"
+            ]
+        }
+    },
+    {
+        "id": 18449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006523.wav"
+            ]
+        }
+    },
+    {
+        "id": 18456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004063.wav"
+            ]
+        }
+    },
+    {
+        "id": 18463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bagpipe makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000693.wav"
+            ]
+        }
+    },
+    {
+        "id": 18465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006254.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006254.wav"
+            ]
+        }
+    },
+    {
+        "id": 18470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003944.wav"
+            ]
+        }
+    },
+    {
+        "id": 18475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006377.wav"
+            ]
+        }
+    },
+    {
+        "id": 18479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005267.wav"
+            ]
+        }
+    },
+    {
+        "id": 18480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002225.wav"
+            ]
+        }
+    },
+    {
+        "id": 18484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001217.wav"
+            ]
+        }
+    },
+    {
+        "id": 18498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003866.wav"
+            ]
+        }
+    },
+    {
+        "id": 18500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008476.wav"
+            ]
+        }
+    },
+    {
+        "id": 18503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008171.wav"
+            ]
+        }
+    },
+    {
+        "id": 18510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001828.wav"
+            ]
+        }
+    },
+    {
+        "id": 18514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000195.wav"
+            ]
+        }
+    },
+    {
+        "id": 18517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003247.wav"
+            ]
+        }
+    },
+    {
+        "id": 18522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005445.wav"
+            ]
+        }
+    },
+    {
+        "id": 18524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006277.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006277.wav"
+            ]
+        }
+    },
+    {
+        "id": 18525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004343.wav"
+            ]
+        }
+    },
+    {
+        "id": 18536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002815.wav"
+            ]
+        }
+    },
+    {
+        "id": 18548,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003682.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003682.wav"
+            ]
+        }
+    },
+    {
+        "id": 18552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007034.wav"
+            ]
+        }
+    },
+    {
+        "id": 18557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many saxophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001972.wav"
+            ]
+        }
+    },
+    {
+        "id": 18564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005857.wav"
+            ]
+        }
+    },
+    {
+        "id": 18568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005067.wav"
+            ]
+        }
+    },
+    {
+        "id": 18569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007626.wav"
+            ]
+        }
+    },
+    {
+        "id": 18571,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001622.wav"
+            ]
+        }
+    },
+    {
+        "id": 18590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005728.wav"
+            ]
+        }
+    },
+    {
+        "id": 18598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002627.wav"
+            ]
+        }
+    },
+    {
+        "id": 18613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002407.wav"
+            ]
+        }
+    },
+    {
+        "id": 18614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006228.wav"
+            ]
+        }
+    },
+    {
+        "id": 18622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion more rhythmic than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005412.wav"
+            ]
+        }
+    },
+    {
+        "id": 18628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006838.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006838.wav"
+            ]
+        }
+    },
+    {
+        "id": 18629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001676.wav"
+            ]
+        }
+    },
+    {
+        "id": 18633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005263.wav"
+            ]
+        }
+    },
+    {
+        "id": 18635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006045.wav"
+            ]
+        }
+    },
+    {
+        "id": 18638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002293.wav"
+            ]
+        }
+    },
+    {
+        "id": 18639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001960.wav"
+            ]
+        }
+    },
+    {
+        "id": 18658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004692.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004692.wav"
+            ]
+        }
+    },
+    {
+        "id": 18662,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007336.wav"
+            ]
+        }
+    },
+    {
+        "id": 18665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004097.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004097.wav"
+            ]
+        }
+    },
+    {
+        "id": 18666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000780.wav"
+            ]
+        }
+    },
+    {
+        "id": 18670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002023.wav"
+            ]
+        }
+    },
+    {
+        "id": 18679,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000970.wav"
+            ]
+        }
+    },
+    {
+        "id": 18684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007513.wav"
+            ]
+        }
+    },
+    {
+        "id": 18699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004186.wav"
+            ]
+        }
+    },
+    {
+        "id": 18700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006927.wav"
+            ]
+        }
+    },
+    {
+        "id": 18706,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004244.wav"
+            ]
+        }
+    },
+    {
+        "id": 18707,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000358.wav"
+            ]
+        }
+    },
+    {
+        "id": 18710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006887.wav"
+            ]
+        }
+    },
+    {
+        "id": 18714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006064.wav"
+            ]
+        }
+    },
+    {
+        "id": 18718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006662.wav"
+            ]
+        }
+    },
+    {
+        "id": 18719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006979.wav"
+            ]
+        }
+    },
+    {
+        "id": 18729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001377.wav"
+            ]
+        }
+    },
+    {
+        "id": 18732,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002901.wav"
+            ]
+        }
+    },
+    {
+        "id": 18734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004549.wav"
+            ]
+        }
+    },
+    {
+        "id": 18735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004327.wav"
+            ]
+        }
+    },
+    {
+        "id": 18745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008885.wav"
+            ]
+        }
+    },
+    {
+        "id": 18756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001887.wav"
+            ]
+        }
+    },
+    {
+        "id": 18759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005119.wav"
+            ]
+        }
+    },
+    {
+        "id": 18765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004898.wav"
+            ]
+        }
+    },
+    {
+        "id": 18770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006211.wav"
+            ]
+        }
+    },
+    {
+        "id": 18780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005651.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005651.wav"
+            ]
+        }
+    },
+    {
+        "id": 18783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000686.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000686.wav"
+            ]
+        }
+    },
+    {
+        "id": 18784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008571.wav"
+            ]
+        }
+    },
+    {
+        "id": 18801,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002595.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002595.wav"
+            ]
+        }
+    },
+    {
+        "id": 18809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003048.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003048.wav"
+            ]
+        }
+    },
+    {
+        "id": 18816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001738.wav"
+            ]
+        }
+    },
+    {
+        "id": 18817,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006882.wav"
+            ]
+        }
+    },
+    {
+        "id": 18820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007880.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007880.wav"
+            ]
+        }
+    },
+    {
+        "id": 18827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008947.wav"
+            ]
+        }
+    },
+    {
+        "id": 18828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006450.wav"
+            ]
+        }
+    },
+    {
+        "id": 18832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003587.wav"
+            ]
+        }
+    },
+    {
+        "id": 18835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006414.wav"
+            ]
+        }
+    },
+    {
+        "id": 18846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000864.wav"
+            ]
+        }
+    },
+    {
+        "id": 18851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008761.wav"
+            ]
+        }
+    },
+    {
+        "id": 18858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003775.wav"
+            ]
+        }
+    },
+    {
+        "id": 18879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008736.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008736.wav"
+            ]
+        }
+    },
+    {
+        "id": 18881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001725.wav"
+            ]
+        }
+    },
+    {
+        "id": 18883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000553.wav"
+            ]
+        }
+    },
+    {
+        "id": 18893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002631.wav"
+            ]
+        }
+    },
+    {
+        "id": 18896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008637.wav"
+            ]
+        }
+    },
+    {
+        "id": 18903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000982.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000982.wav"
+            ]
+        }
+    },
+    {
+        "id": 18909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007440.wav"
+            ]
+        }
+    },
+    {
+        "id": 18912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007683.wav"
+            ]
+        }
+    },
+    {
+        "id": 18930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007981.wav"
+            ]
+        }
+    },
+    {
+        "id": 18938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005488.wav"
+            ]
+        }
+    },
+    {
+        "id": 18942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007869.wav"
+            ]
+        }
+    },
+    {
+        "id": 18943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004630.wav"
+            ]
+        }
+    },
+    {
+        "id": 18944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002910.wav"
+            ]
+        }
+    },
+    {
+        "id": 18945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many electric_bass are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005828.wav"
+            ]
+        }
+    },
+    {
+        "id": 18946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005789.wav"
+            ]
+        }
+    },
+    {
+        "id": 18950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003908.wav"
+            ]
+        }
+    },
+    {
+        "id": 18951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004522.wav"
+            ]
+        }
+    },
+    {
+        "id": 18952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005974.wav"
+            ]
+        }
+    },
+    {
+        "id": 18956,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008240.wav"
+            ]
+        }
+    },
+    {
+        "id": 18969,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008510.wav"
+            ]
+        }
+    },
+    {
+        "id": 18987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007914.wav"
+            ]
+        }
+    },
+    {
+        "id": 18995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001426.wav"
+            ]
+        }
+    },
+    {
+        "id": 18997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007504.wav"
+            ]
+        }
+    },
+    {
+        "id": 19010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003709.wav"
+            ]
+        }
+    },
+    {
+        "id": 19017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000429.wav"
+            ]
+        }
+    },
+    {
+        "id": 19021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002731.wav"
+            ]
+        }
+    },
+    {
+        "id": 19022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003772.wav"
+            ]
+        }
+    },
+    {
+        "id": 19035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007708.wav"
+            ]
+        }
+    },
+    {
+        "id": 19038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005230.wav"
+            ]
+        }
+    },
+    {
+        "id": 19055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002696.wav"
+            ]
+        }
+    },
+    {
+        "id": 19067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006636.wav"
+            ]
+        }
+    },
+    {
+        "id": 19072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004297.wav"
+            ]
+        }
+    },
+    {
+        "id": 19073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008223.wav"
+            ]
+        }
+    },
+    {
+        "id": 19077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007309.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007309.wav"
+            ]
+        }
+    },
+    {
+        "id": 19080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001503.wav"
+            ]
+        }
+    },
+    {
+        "id": 19085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007789.wav"
+            ]
+        }
+    },
+    {
+        "id": 19086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000222.wav"
+            ]
+        }
+    },
+    {
+        "id": 19088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001336.wav"
+            ]
+        }
+    },
+    {
+        "id": 19094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002915.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002915.wav"
+            ]
+        }
+    },
+    {
+        "id": 19101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003623.wav"
+            ]
+        }
+    },
+    {
+        "id": 19102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003455.wav"
+            ]
+        }
+    },
+    {
+        "id": 19103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007896.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007896.wav"
+            ]
+        }
+    },
+    {
+        "id": 19105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007903.wav"
+            ]
+        }
+    },
+    {
+        "id": 19106,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006108.wav"
+            ]
+        }
+    },
+    {
+        "id": 19107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005667.wav"
+            ]
+        }
+    },
+    {
+        "id": 19111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005331.wav"
+            ]
+        }
+    },
+    {
+        "id": 19115,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007290.wav"
+            ]
+        }
+    },
+    {
+        "id": 19125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003459.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003459.wav"
+            ]
+        }
+    },
+    {
+        "id": 19136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006514.wav"
+            ]
+        }
+    },
+    {
+        "id": 19139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005835.wav"
+            ]
+        }
+    },
+    {
+        "id": 19141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006071.wav"
+            ]
+        }
+    },
+    {
+        "id": 19142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008957.wav"
+            ]
+        }
+    },
+    {
+        "id": 19143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008546.wav"
+            ]
+        }
+    },
+    {
+        "id": 19150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008881.wav"
+            ]
+        }
+    },
+    {
+        "id": 19152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000967.wav"
+            ]
+        }
+    },
+    {
+        "id": 19155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001100.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001100.wav"
+            ]
+        }
+    },
+    {
+        "id": 19162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003427.wav"
+            ]
+        }
+    },
+    {
+        "id": 19168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001218.wav"
+            ]
+        }
+    },
+    {
+        "id": 19175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005376.wav"
+            ]
+        }
+    },
+    {
+        "id": 19180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008564.wav"
+            ]
+        }
+    },
+    {
+        "id": 19212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003904.wav"
+            ]
+        }
+    },
+    {
+        "id": 19214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007212.wav"
+            ]
+        }
+    },
+    {
+        "id": 19217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003856.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003856.wav"
+            ]
+        }
+    },
+    {
+        "id": 19219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and congas instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007458.wav"
+            ]
+        }
+    },
+    {
+        "id": 19220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006680.wav"
+            ]
+        }
+    },
+    {
+        "id": 19228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006903.wav"
+            ]
+        }
+    },
+    {
+        "id": 19229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002777.wav"
+            ]
+        }
+    },
+    {
+        "id": 19232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007876.wav"
+            ]
+        }
+    },
+    {
+        "id": 19233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007411.wav"
+            ]
+        }
+    },
+    {
+        "id": 19234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004772.wav"
+            ]
+        }
+    },
+    {
+        "id": 19235,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001882.wav"
+            ]
+        }
+    },
+    {
+        "id": 19237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007369.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007369.wav"
+            ]
+        }
+    },
+    {
+        "id": 19253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005779.wav"
+            ]
+        }
+    },
+    {
+        "id": 19255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001043.wav"
+            ]
+        }
+    },
+    {
+        "id": 19265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006774.wav"
+            ]
+        }
+    },
+    {
+        "id": 19268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007904.wav"
+            ]
+        }
+    },
+    {
+        "id": 19271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006797.wav"
+            ]
+        }
+    },
+    {
+        "id": 19276,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000388.wav"
+            ]
+        }
+    },
+    {
+        "id": 19277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the clarinet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000984.wav"
+            ]
+        }
+    },
+    {
+        "id": 19281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001311.wav"
+            ]
+        }
+    },
+    {
+        "id": 19304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002326.wav"
+            ]
+        }
+    },
+    {
+        "id": 19317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003147.wav"
+            ]
+        }
+    },
+    {
+        "id": 19319,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005362.wav"
+            ]
+        }
+    },
+    {
+        "id": 19327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006897.wav"
+            ]
+        }
+    },
+    {
+        "id": 19335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002947.wav"
+            ]
+        }
+    },
+    {
+        "id": 19350,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007639.wav"
+            ]
+        }
+    },
+    {
+        "id": 19351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002332.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002332.wav"
+            ]
+        }
+    },
+    {
+        "id": 19352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005689.wav"
+            ]
+        }
+    },
+    {
+        "id": 19353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000269.wav"
+            ]
+        }
+    },
+    {
+        "id": 19358,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich xylophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004050.wav"
+            ]
+        }
+    },
+    {
+        "id": 19364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008612.wav"
+            ]
+        }
+    },
+    {
+        "id": 19369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001189.wav"
+            ]
+        }
+    },
+    {
+        "id": 19374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004012.wav"
+            ]
+        }
+    },
+    {
+        "id": 19375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008419.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008419.wav"
+            ]
+        }
+    },
+    {
+        "id": 19376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008870.wav"
+            ]
+        }
+    },
+    {
+        "id": 19384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000704.wav"
+            ]
+        }
+    },
+    {
+        "id": 19418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007563.wav"
+            ]
+        }
+    },
+    {
+        "id": 19420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002637.wav"
+            ]
+        }
+    },
+    {
+        "id": 19432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004218.wav"
+            ]
+        }
+    },
+    {
+        "id": 19438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002629.wav"
+            ]
+        }
+    },
+    {
+        "id": 19447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006450.wav"
+            ]
+        }
+    },
+    {
+        "id": 19450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008098.wav"
+            ]
+        }
+    },
+    {
+        "id": 19454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001819.wav"
+            ]
+        }
+    },
+    {
+        "id": 19458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005865.wav"
+            ]
+        }
+    },
+    {
+        "id": 19462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003388.wav"
+            ]
+        }
+    },
+    {
+        "id": 19471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001762.wav"
+            ]
+        }
+    },
+    {
+        "id": 19479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006548.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006548.wav"
+            ]
+        }
+    },
+    {
+        "id": 19486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001620.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001620.wav"
+            ]
+        }
+    },
+    {
+        "id": 19487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007615.wav"
+            ]
+        }
+    },
+    {
+        "id": 19490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003739.wav"
+            ]
+        }
+    },
+    {
+        "id": 19495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003970.wav"
+            ]
+        }
+    },
+    {
+        "id": 19502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001774.wav"
+            ]
+        }
+    },
+    {
+        "id": 19521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002582.wav"
+            ]
+        }
+    },
+    {
+        "id": 19526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007559.wav"
+            ]
+        }
+    },
+    {
+        "id": 19533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006018.wav"
+            ]
+        }
+    },
+    {
+        "id": 19539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005066.wav"
+            ]
+        }
+    },
+    {
+        "id": 19546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004054.wav"
+            ]
+        }
+    },
+    {
+        "id": 19549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004116.wav"
+            ]
+        }
+    },
+    {
+        "id": 19561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000204.wav"
+            ]
+        }
+    },
+    {
+        "id": 19563,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008476.wav"
+            ]
+        }
+    },
+    {
+        "id": 19571,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004228.wav"
+            ]
+        }
+    },
+    {
+        "id": 19572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001382.wav"
+            ]
+        }
+    },
+    {
+        "id": 19574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005247.wav"
+            ]
+        }
+    },
+    {
+        "id": 19575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008633.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008633.wav"
+            ]
+        }
+    },
+    {
+        "id": 19576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004345.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004345.wav"
+            ]
+        }
+    },
+    {
+        "id": 19577,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008063.wav"
+            ]
+        }
+    },
+    {
+        "id": 19580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007300.wav"
+            ]
+        }
+    },
+    {
+        "id": 19584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the right more rhythmic than the guzheng on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006387.wav"
+            ]
+        }
+    },
+    {
+        "id": 19587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004172.wav"
+            ]
+        }
+    },
+    {
+        "id": 19589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005550.wav"
+            ]
+        }
+    },
+    {
+        "id": 19604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006412.wav"
+            ]
+        }
+    },
+    {
+        "id": 19605,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008228.wav"
+            ]
+        }
+    },
+    {
+        "id": 19612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004414.wav"
+            ]
+        }
+    },
+    {
+        "id": 19613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001825.wav"
+            ]
+        }
+    },
+    {
+        "id": 19637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005054.wav"
+            ]
+        }
+    },
+    {
+        "id": 19639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007591.wav"
+            ]
+        }
+    },
+    {
+        "id": 19650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000601.wav"
+            ]
+        }
+    },
+    {
+        "id": 19653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002740.wav"
+            ]
+        }
+    },
+    {
+        "id": 19654,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004658.wav"
+            ]
+        }
+    },
+    {
+        "id": 19655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002937.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002937.wav"
+            ]
+        }
+    },
+    {
+        "id": 19660,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000509.wav"
+            ]
+        }
+    },
+    {
+        "id": 19661,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005675.wav"
+            ]
+        }
+    },
+    {
+        "id": 19662,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006977.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006977.wav"
+            ]
+        }
+    },
+    {
+        "id": 19668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001525.wav"
+            ]
+        }
+    },
+    {
+        "id": 19669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006644.wav"
+            ]
+        }
+    },
+    {
+        "id": 19677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003120.wav"
+            ]
+        }
+    },
+    {
+        "id": 19678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001361.wav"
+            ]
+        }
+    },
+    {
+        "id": 19679,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003142.wav"
+            ]
+        }
+    },
+    {
+        "id": 19683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000739.wav"
+            ]
+        }
+    },
+    {
+        "id": 19689,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008628.wav"
+            ]
+        }
+    },
+    {
+        "id": 19693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002558.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002558.wav"
+            ]
+        }
+    },
+    {
+        "id": 19694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008182.wav"
+            ]
+        }
+    },
+    {
+        "id": 19695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000600.wav"
+            ]
+        }
+    },
+    {
+        "id": 19699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002535.wav"
+            ]
+        }
+    },
+    {
+        "id": 19700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008507.wav"
+            ]
+        }
+    },
+    {
+        "id": 19702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008026.wav"
+            ]
+        }
+    },
+    {
+        "id": 19704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000645.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000645.wav"
+            ]
+        }
+    },
+    {
+        "id": 19723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008567.wav"
+            ]
+        }
+    },
+    {
+        "id": 19731,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006960.wav"
+            ]
+        }
+    },
+    {
+        "id": 19733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008385.wav"
+            ]
+        }
+    },
+    {
+        "id": 19748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000402.wav"
+            ]
+        }
+    },
+    {
+        "id": 19763,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006094.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006094.wav"
+            ]
+        }
+    },
+    {
+        "id": 19777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003546.wav"
+            ]
+        }
+    },
+    {
+        "id": 19790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000536.wav"
+            ]
+        }
+    },
+    {
+        "id": 19798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000162.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000162.wav"
+            ]
+        }
+    },
+    {
+        "id": 19804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003347.wav"
+            ]
+        }
+    },
+    {
+        "id": 19809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006662.wav"
+            ]
+        }
+    },
+    {
+        "id": 19814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008872.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008872.wav"
+            ]
+        }
+    },
+    {
+        "id": 19816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006243.wav"
+            ]
+        }
+    },
+    {
+        "id": 19821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006423.wav"
+            ]
+        }
+    },
+    {
+        "id": 19828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005984.wav"
+            ]
+        }
+    },
+    {
+        "id": 19832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003116.wav"
+            ]
+        }
+    },
+    {
+        "id": 19834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004375.wav"
+            ]
+        }
+    },
+    {
+        "id": 19842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003763.wav"
+            ]
+        }
+    },
+    {
+        "id": 19847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002790.wav"
+            ]
+        }
+    },
+    {
+        "id": 19858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006466.wav"
+            ]
+        }
+    },
+    {
+        "id": 19862,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006051.wav"
+            ]
+        }
+    },
+    {
+        "id": 19869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000061.wav"
+            ]
+        }
+    },
+    {
+        "id": 19873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002672.wav"
+            ]
+        }
+    },
+    {
+        "id": 19880,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008457.wav"
+            ]
+        }
+    },
+    {
+        "id": 19883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006687.wav"
+            ]
+        }
+    },
+    {
+        "id": 19887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000245.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000245.wav"
+            ]
+        }
+    },
+    {
+        "id": 19889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002211.wav"
+            ]
+        }
+    },
+    {
+        "id": 19892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004961.wav"
+            ]
+        }
+    },
+    {
+        "id": 19899,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001469.wav"
+            ]
+        }
+    },
+    {
+        "id": 19916,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002317.wav"
+            ]
+        }
+    },
+    {
+        "id": 19917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007021.wav"
+            ]
+        }
+    },
+    {
+        "id": 19919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001297.wav"
+            ]
+        }
+    },
+    {
+        "id": 19920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008925.wav"
+            ]
+        }
+    },
+    {
+        "id": 19921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002979.wav"
+            ]
+        }
+    },
+    {
+        "id": 19922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003916.wav"
+            ]
+        }
+    },
+    {
+        "id": 19924,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000034.wav"
+            ]
+        }
+    },
+    {
+        "id": 19928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008299.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008299.wav"
+            ]
+        }
+    },
+    {
+        "id": 19930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008858.wav"
+            ]
+        }
+    },
+    {
+        "id": 19936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008056.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008056.wav"
+            ]
+        }
+    },
+    {
+        "id": 19937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000495.wav"
+            ]
+        }
+    },
+    {
+        "id": 19940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007678.wav"
+            ]
+        }
+    },
+    {
+        "id": 19946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000914.wav"
+            ]
+        }
+    },
+    {
+        "id": 19948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005719.wav"
+            ]
+        }
+    },
+    {
+        "id": 19957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many trumpet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001944.wav"
+            ]
+        }
+    },
+    {
+        "id": 19967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008973.wav"
+            ]
+        }
+    },
+    {
+        "id": 19974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008317.wav"
+            ]
+        }
+    },
+    {
+        "id": 19976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001661.wav"
+            ]
+        }
+    },
+    {
+        "id": 19980,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008308.wav"
+            ]
+        }
+    },
+    {
+        "id": 19985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004316.wav"
+            ]
+        }
+    },
+    {
+        "id": 20010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003201.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003201.wav"
+            ]
+        }
+    },
+    {
+        "id": 20012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002505.wav"
+            ]
+        }
+    },
+    {
+        "id": 20029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008945.wav"
+            ]
+        }
+    },
+    {
+        "id": 20033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002848.wav"
+            ]
+        }
+    },
+    {
+        "id": 20034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005829.wav"
+            ]
+        }
+    },
+    {
+        "id": 20036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000677.wav"
+            ]
+        }
+    },
+    {
+        "id": 20041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005147.wav"
+            ]
+        }
+    },
+    {
+        "id": 20043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006470.wav"
+            ]
+        }
+    },
+    {
+        "id": 20049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004471.wav"
+            ]
+        }
+    },
+    {
+        "id": 20057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007378.wav"
+            ]
+        }
+    },
+    {
+        "id": 20059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002806.wav"
+            ]
+        }
+    },
+    {
+        "id": 20060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006389.wav"
+            ]
+        }
+    },
+    {
+        "id": 20071,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008662.wav"
+            ]
+        }
+    },
+    {
+        "id": 20076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001598.wav"
+            ]
+        }
+    },
+    {
+        "id": 20079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003564.wav"
+            ]
+        }
+    },
+    {
+        "id": 20081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007800.wav"
+            ]
+        }
+    },
+    {
+        "id": 20086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005033.wav"
+            ]
+        }
+    },
+    {
+        "id": 20099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008152.wav"
+            ]
+        }
+    },
+    {
+        "id": 20102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004427.wav"
+            ]
+        }
+    },
+    {
+        "id": 20104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004756.wav"
+            ]
+        }
+    },
+    {
+        "id": 20107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007553.wav"
+            ]
+        }
+    },
+    {
+        "id": 20114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005527.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005527.wav"
+            ]
+        }
+    },
+    {
+        "id": 20116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000065.wav"
+            ]
+        }
+    },
+    {
+        "id": 20126,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005376.wav"
+            ]
+        }
+    },
+    {
+        "id": 20131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004847.wav"
+            ]
+        }
+    },
+    {
+        "id": 20133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007011.wav"
+            ]
+        }
+    },
+    {
+        "id": 20138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007467.wav"
+            ]
+        }
+    },
+    {
+        "id": 20139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002444.wav"
+            ]
+        }
+    },
+    {
+        "id": 20149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007294.wav"
+            ]
+        }
+    },
+    {
+        "id": 20150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000641.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000641.wav"
+            ]
+        }
+    },
+    {
+        "id": 20162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005841.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005841.wav"
+            ]
+        }
+    },
+    {
+        "id": 20163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004251.wav"
+            ]
+        }
+    },
+    {
+        "id": 20164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004380.wav"
+            ]
+        }
+    },
+    {
+        "id": 20175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001640.wav"
+            ]
+        }
+    },
+    {
+        "id": 20181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007506.wav"
+            ]
+        }
+    },
+    {
+        "id": 20187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002420.wav"
+            ]
+        }
+    },
+    {
+        "id": 20189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004094.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004094.wav"
+            ]
+        }
+    },
+    {
+        "id": 20190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003761.wav"
+            ]
+        }
+    },
+    {
+        "id": 20193,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006175.wav"
+            ]
+        }
+    },
+    {
+        "id": 20198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001468.wav"
+            ]
+        }
+    },
+    {
+        "id": 20208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001602.wav"
+            ]
+        }
+    },
+    {
+        "id": 20217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007389.wav"
+            ]
+        }
+    },
+    {
+        "id": 20221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006117.wav"
+            ]
+        }
+    },
+    {
+        "id": 20224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008397.wav"
+            ]
+        }
+    },
+    {
+        "id": 20228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001780.wav"
+            ]
+        }
+    },
+    {
+        "id": 20231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003287.wav"
+            ]
+        }
+    },
+    {
+        "id": 20234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005806.wav"
+            ]
+        }
+    },
+    {
+        "id": 20242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004457.wav"
+            ]
+        }
+    },
+    {
+        "id": 20243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003377.wav"
+            ]
+        }
+    },
+    {
+        "id": 20249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007200.wav"
+            ]
+        }
+    },
+    {
+        "id": 20256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005481.wav"
+            ]
+        }
+    },
+    {
+        "id": 20273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007247.wav"
+            ]
+        }
+    },
+    {
+        "id": 20279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003217.wav"
+            ]
+        }
+    },
+    {
+        "id": 20297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004476.wav"
+            ]
+        }
+    },
+    {
+        "id": 20299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006858.wav"
+            ]
+        }
+    },
+    {
+        "id": 20302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003297.wav"
+            ]
+        }
+    },
+    {
+        "id": 20303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001166.wav"
+            ]
+        }
+    },
+    {
+        "id": 20307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008454.wav"
+            ]
+        }
+    },
+    {
+        "id": 20312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005241.wav"
+            ]
+        }
+    },
+    {
+        "id": 20314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000735.wav"
+            ]
+        }
+    },
+    {
+        "id": 20315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007287.wav"
+            ]
+        }
+    },
+    {
+        "id": 20317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003421.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003421.wav"
+            ]
+        }
+    },
+    {
+        "id": 20325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002704.wav"
+            ]
+        }
+    },
+    {
+        "id": 20330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001095.wav"
+            ]
+        }
+    },
+    {
+        "id": 20331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004250.wav"
+            ]
+        }
+    },
+    {
+        "id": 20333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000621.wav"
+            ]
+        }
+    },
+    {
+        "id": 20335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006379.wav"
+            ]
+        }
+    },
+    {
+        "id": 20344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004771.wav"
+            ]
+        }
+    },
+    {
+        "id": 20346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002442.wav"
+            ]
+        }
+    },
+    {
+        "id": 20349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002437.wav"
+            ]
+        }
+    },
+    {
+        "id": 20351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006782.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006782.wav"
+            ]
+        }
+    },
+    {
+        "id": 20365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006386.wav"
+            ]
+        }
+    },
+    {
+        "id": 20375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003392.wav"
+            ]
+        }
+    },
+    {
+        "id": 20379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008058.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008058.wav"
+            ]
+        }
+    },
+    {
+        "id": 20382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002773.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002773.wav"
+            ]
+        }
+    },
+    {
+        "id": 20384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005456.wav"
+            ]
+        }
+    },
+    {
+        "id": 20391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002531.wav"
+            ]
+        }
+    },
+    {
+        "id": 20396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007074.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007074.wav"
+            ]
+        }
+    },
+    {
+        "id": 20405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000933.wav"
+            ]
+        }
+    },
+    {
+        "id": 20407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000436.wav"
+            ]
+        }
+    },
+    {
+        "id": 20414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002030.wav"
+            ]
+        }
+    },
+    {
+        "id": 20422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006798.wav"
+            ]
+        }
+    },
+    {
+        "id": 20427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003528.wav"
+            ]
+        }
+    },
+    {
+        "id": 20428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006513.wav"
+            ]
+        }
+    },
+    {
+        "id": 20431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005778.wav"
+            ]
+        }
+    },
+    {
+        "id": 20440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001204.wav"
+            ]
+        }
+    },
+    {
+        "id": 20450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003717.wav"
+            ]
+        }
+    },
+    {
+        "id": 20455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004782.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004782.wav"
+            ]
+        }
+    },
+    {
+        "id": 20457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000844.wav"
+            ]
+        }
+    },
+    {
+        "id": 20458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000890.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000890.wav"
+            ]
+        }
+    },
+    {
+        "id": 20460,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002550.wav"
+            ]
+        }
+    },
+    {
+        "id": 20461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005589.wav"
+            ]
+        }
+    },
+    {
+        "id": 20474,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007775.wav"
+            ]
+        }
+    },
+    {
+        "id": 20477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004597.wav"
+            ]
+        }
+    },
+    {
+        "id": 20485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many congas are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001780.wav"
+            ]
+        }
+    },
+    {
+        "id": 20487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001744.wav"
+            ]
+        }
+    },
+    {
+        "id": 20499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000583.wav"
+            ]
+        }
+    },
+    {
+        "id": 20500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007209.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007209.wav"
+            ]
+        }
+    },
+    {
+        "id": 20503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001746.wav"
+            ]
+        }
+    },
+    {
+        "id": 20504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001023.wav"
+            ]
+        }
+    },
+    {
+        "id": 20513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007235.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007235.wav"
+            ]
+        }
+    },
+    {
+        "id": 20519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002411.wav"
+            ]
+        }
+    },
+    {
+        "id": 20522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004550.wav"
+            ]
+        }
+    },
+    {
+        "id": 20530,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006099.wav"
+            ]
+        }
+    },
+    {
+        "id": 20538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007530.wav"
+            ]
+        }
+    },
+    {
+        "id": 20542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002481.wav"
+            ]
+        }
+    },
+    {
+        "id": 20543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008168.wav"
+            ]
+        }
+    },
+    {
+        "id": 20547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001389.wav"
+            ]
+        }
+    },
+    {
+        "id": 20549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002974.wav"
+            ]
+        }
+    },
+    {
+        "id": 20555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007040.wav"
+            ]
+        }
+    },
+    {
+        "id": 20559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001948.wav"
+            ]
+        }
+    },
+    {
+        "id": 20560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002035.wav"
+            ]
+        }
+    },
+    {
+        "id": 20567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001401.wav"
+            ]
+        }
+    },
+    {
+        "id": 20574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007949.wav"
+            ]
+        }
+    },
+    {
+        "id": 20575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004485.wav"
+            ]
+        }
+    },
+    {
+        "id": 20580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003221.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003221.wav"
+            ]
+        }
+    },
+    {
+        "id": 20581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005628.wav"
+            ]
+        }
+    },
+    {
+        "id": 20588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000101.wav"
+            ]
+        }
+    },
+    {
+        "id": 20590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000748.wav"
+            ]
+        }
+    },
+    {
+        "id": 20596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004804.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004804.wav"
+            ]
+        }
+    },
+    {
+        "id": 20598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006398.wav"
+            ]
+        }
+    },
+    {
+        "id": 20599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001500.wav"
+            ]
+        }
+    },
+    {
+        "id": 20602,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003364.wav"
+            ]
+        }
+    },
+    {
+        "id": 20612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004703.wav"
+            ]
+        }
+    },
+    {
+        "id": 20624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004276.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004276.wav"
+            ]
+        }
+    },
+    {
+        "id": 20628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000758.wav"
+            ]
+        }
+    },
+    {
+        "id": 20634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004184.wav"
+            ]
+        }
+    },
+    {
+        "id": 20642,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005717.wav"
+            ]
+        }
+    },
+    {
+        "id": 20649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006486.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006486.wav"
+            ]
+        }
+    },
+    {
+        "id": 20656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008351.wav"
+            ]
+        }
+    },
+    {
+        "id": 20668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007150.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007150.wav"
+            ]
+        }
+    },
+    {
+        "id": 20670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a congas in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008232.wav"
+            ]
+        }
+    },
+    {
+        "id": 20682,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002200.wav"
+            ]
+        }
+    },
+    {
+        "id": 20684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001515.wav"
+            ]
+        }
+    },
+    {
+        "id": 20685,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007874.wav"
+            ]
+        }
+    },
+    {
+        "id": 20694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002816.wav"
+            ]
+        }
+    },
+    {
+        "id": 20698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many drum are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000212.wav"
+            ]
+        }
+    },
+    {
+        "id": 20706,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003234.wav"
+            ]
+        }
+    },
+    {
+        "id": 20707,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006380.wav"
+            ]
+        }
+    },
+    {
+        "id": 20710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006271.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006271.wav"
+            ]
+        }
+    },
+    {
+        "id": 20714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004724.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004724.wav"
+            ]
+        }
+    },
+    {
+        "id": 20718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006581.wav"
+            ]
+        }
+    },
+    {
+        "id": 20728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000037.wav"
+            ]
+        }
+    },
+    {
+        "id": 20729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007961.wav"
+            ]
+        }
+    },
+    {
+        "id": 20730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004229.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004229.wav"
+            ]
+        }
+    },
+    {
+        "id": 20733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004354.wav"
+            ]
+        }
+    },
+    {
+        "id": 20736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many pipa are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004722.wav"
+            ]
+        }
+    },
+    {
+        "id": 20737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000485.wav"
+            ]
+        }
+    },
+    {
+        "id": 20748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004510.wav"
+            ]
+        }
+    },
+    {
+        "id": 20750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003897.wav"
+            ]
+        }
+    },
+    {
+        "id": 20757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008899.wav"
+            ]
+        }
+    },
+    {
+        "id": 20758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003526.wav"
+            ]
+        }
+    },
+    {
+        "id": 20762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006702.wav"
+            ]
+        }
+    },
+    {
+        "id": 20764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004810.wav"
+            ]
+        }
+    },
+    {
+        "id": 20766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006874.wav"
+            ]
+        }
+    },
+    {
+        "id": 20769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008015.wav"
+            ]
+        }
+    },
+    {
+        "id": 20770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008784.wav"
+            ]
+        }
+    },
+    {
+        "id": 20772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004834.wav"
+            ]
+        }
+    },
+    {
+        "id": 20774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000450.wav"
+            ]
+        }
+    },
+    {
+        "id": 20778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005045.wav"
+            ]
+        }
+    },
+    {
+        "id": 20785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006488.wav"
+            ]
+        }
+    },
+    {
+        "id": 20786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008308.wav"
+            ]
+        }
+    },
+    {
+        "id": 20787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007247.wav"
+            ]
+        }
+    },
+    {
+        "id": 20791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003199.wav"
+            ]
+        }
+    },
+    {
+        "id": 20795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005885.wav"
+            ]
+        }
+    },
+    {
+        "id": 20796,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001734.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001734.wav"
+            ]
+        }
+    },
+    {
+        "id": 20800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001559.wav"
+            ]
+        }
+    },
+    {
+        "id": 20803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002835.wav"
+            ]
+        }
+    },
+    {
+        "id": 20804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007227.wav"
+            ]
+        }
+    },
+    {
+        "id": 20816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004611.wav"
+            ]
+        }
+    },
+    {
+        "id": 20825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many flute are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000562.wav"
+            ]
+        }
+    },
+    {
+        "id": 20828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006445.wav"
+            ]
+        }
+    },
+    {
+        "id": 20831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008655.wav"
+            ]
+        }
+    },
+    {
+        "id": 20832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008711.wav"
+            ]
+        }
+    },
+    {
+        "id": 20838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000995.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000995.wav"
+            ]
+        }
+    },
+    {
+        "id": 20839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008389.wav"
+            ]
+        }
+    },
+    {
+        "id": 20851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007597.wav"
+            ]
+        }
+    },
+    {
+        "id": 20853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006279.wav"
+            ]
+        }
+    },
+    {
+        "id": 20854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003659.wav"
+            ]
+        }
+    },
+    {
+        "id": 20855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004846.wav"
+            ]
+        }
+    },
+    {
+        "id": 20856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004112.wav"
+            ]
+        }
+    },
+    {
+        "id": 20862,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005218.wav"
+            ]
+        }
+    },
+    {
+        "id": 20865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000728.wav"
+            ]
+        }
+    },
+    {
+        "id": 20868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003212.wav"
+            ]
+        }
+    },
+    {
+        "id": 20871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006593.wav"
+            ]
+        }
+    },
+    {
+        "id": 20873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000134.wav"
+            ]
+        }
+    },
+    {
+        "id": 20883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002294.wav"
+            ]
+        }
+    },
+    {
+        "id": 20902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002659.wav"
+            ]
+        }
+    },
+    {
+        "id": 20911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004951.wav"
+            ]
+        }
+    },
+    {
+        "id": 20912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005754.wav"
+            ]
+        }
+    },
+    {
+        "id": 20915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003281.wav"
+            ]
+        }
+    },
+    {
+        "id": 20928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005007.wav"
+            ]
+        }
+    },
+    {
+        "id": 20934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001082.wav"
+            ]
+        }
+    },
+    {
+        "id": 20940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008357.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008357.wav"
+            ]
+        }
+    },
+    {
+        "id": 20948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007704.wav"
+            ]
+        }
+    },
+    {
+        "id": 20952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002966.wav"
+            ]
+        }
+    },
+    {
+        "id": 20971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008406.wav"
+            ]
+        }
+    },
+    {
+        "id": 20972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003329.wav"
+            ]
+        }
+    },
+    {
+        "id": 20981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006351.wav"
+            ]
+        }
+    },
+    {
+        "id": 20988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002915.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002915.wav"
+            ]
+        }
+    },
+    {
+        "id": 20999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008340.wav"
+            ]
+        }
+    },
+    {
+        "id": 21003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002426.wav"
+            ]
+        }
+    },
+    {
+        "id": 21012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002884.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002884.wav"
+            ]
+        }
+    },
+    {
+        "id": 21013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002235.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002235.wav"
+            ]
+        }
+    },
+    {
+        "id": 21017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a pipa in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003133.wav"
+            ]
+        }
+    },
+    {
+        "id": 21026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many guzheng are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000530.wav"
+            ]
+        }
+    },
+    {
+        "id": 21028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001117.wav"
+            ]
+        }
+    },
+    {
+        "id": 21037,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002256.wav"
+            ]
+        }
+    },
+    {
+        "id": 21042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004383.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004383.wav"
+            ]
+        }
+    },
+    {
+        "id": 21053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002103.wav"
+            ]
+        }
+    },
+    {
+        "id": 21054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001121.wav"
+            ]
+        }
+    },
+    {
+        "id": 21055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many xylophone are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000748.wav"
+            ]
+        }
+    },
+    {
+        "id": 21058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006524.wav"
+            ]
+        }
+    },
+    {
+        "id": 21062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007573.wav"
+            ]
+        }
+    },
+    {
+        "id": 21081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004857.wav"
+            ]
+        }
+    },
+    {
+        "id": 21083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002587.wav"
+            ]
+        }
+    },
+    {
+        "id": 21098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008766.wav"
+            ]
+        }
+    },
+    {
+        "id": 21100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007218.wav"
+            ]
+        }
+    },
+    {
+        "id": 21101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004037.wav"
+            ]
+        }
+    },
+    {
+        "id": 21103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000891.wav"
+            ]
+        }
+    },
+    {
+        "id": 21107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003938.wav"
+            ]
+        }
+    },
+    {
+        "id": 21110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000852.wav"
+            ]
+        }
+    },
+    {
+        "id": 21116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007361.wav"
+            ]
+        }
+    },
+    {
+        "id": 21118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002646.wav"
+            ]
+        }
+    },
+    {
+        "id": 21122,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003590.wav"
+            ]
+        }
+    },
+    {
+        "id": 21125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003993.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003993.wav"
+            ]
+        }
+    },
+    {
+        "id": 21128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002802.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002802.wav"
+            ]
+        }
+    },
+    {
+        "id": 21130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004869.wav"
+            ]
+        }
+    },
+    {
+        "id": 21132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000931.wav"
+            ]
+        }
+    },
+    {
+        "id": 21138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005781.wav"
+            ]
+        }
+    },
+    {
+        "id": 21154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007047.wav"
+            ]
+        }
+    },
+    {
+        "id": 21158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001715.wav"
+            ]
+        }
+    },
+    {
+        "id": 21164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001979.wav"
+            ]
+        }
+    },
+    {
+        "id": 21166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007691.wav"
+            ]
+        }
+    },
+    {
+        "id": 21170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002875.wav"
+            ]
+        }
+    },
+    {
+        "id": 21171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002912.wav"
+            ]
+        }
+    },
+    {
+        "id": 21173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bagpipe are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007807.wav"
+            ]
+        }
+    },
+    {
+        "id": 21174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001327.wav"
+            ]
+        }
+    },
+    {
+        "id": 21179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000103.wav"
+            ]
+        }
+    },
+    {
+        "id": 21192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007597.wav"
+            ]
+        }
+    },
+    {
+        "id": 21205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007487.wav"
+            ]
+        }
+    },
+    {
+        "id": 21207,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000334.wav"
+            ]
+        }
+    },
+    {
+        "id": 21214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005671.wav"
+            ]
+        }
+    },
+    {
+        "id": 21215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008763.wav"
+            ]
+        }
+    },
+    {
+        "id": 21217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007453.wav"
+            ]
+        }
+    },
+    {
+        "id": 21221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002454.wav"
+            ]
+        }
+    },
+    {
+        "id": 21224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003848.wav"
+            ]
+        }
+    },
+    {
+        "id": 21230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006380.wav"
+            ]
+        }
+    },
+    {
+        "id": 21233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007049.wav"
+            ]
+        }
+    },
+    {
+        "id": 21253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000633.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000633.wav"
+            ]
+        }
+    },
+    {
+        "id": 21256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002103.wav"
+            ]
+        }
+    },
+    {
+        "id": 21267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003588.wav"
+            ]
+        }
+    },
+    {
+        "id": 21268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004109.wav"
+            ]
+        }
+    },
+    {
+        "id": 21274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the drum on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004189.wav"
+            ]
+        }
+    },
+    {
+        "id": 21276,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004855.wav"
+            ]
+        }
+    },
+    {
+        "id": 21280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002187.wav"
+            ]
+        }
+    },
+    {
+        "id": 21286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003803.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003803.wav"
+            ]
+        }
+    },
+    {
+        "id": 21301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008437.wav"
+            ]
+        }
+    },
+    {
+        "id": 21309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004044.wav"
+            ]
+        }
+    },
+    {
+        "id": 21310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005464.wav"
+            ]
+        }
+    },
+    {
+        "id": 21321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004696.wav"
+            ]
+        }
+    },
+    {
+        "id": 21322,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many accordion are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006597.wav"
+            ]
+        }
+    },
+    {
+        "id": 21325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001435.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001435.wav"
+            ]
+        }
+    },
+    {
+        "id": 21329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007929.wav"
+            ]
+        }
+    },
+    {
+        "id": 21336,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008251.wav"
+            ]
+        }
+    },
+    {
+        "id": 21338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007035.wav"
+            ]
+        }
+    },
+    {
+        "id": 21340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006084.wav"
+            ]
+        }
+    },
+    {
+        "id": 21342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006887.wav"
+            ]
+        }
+    },
+    {
+        "id": 21345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001680.wav"
+            ]
+        }
+    },
+    {
+        "id": 21354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001395.wav"
+            ]
+        }
+    },
+    {
+        "id": 21355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003745.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003745.wav"
+            ]
+        }
+    },
+    {
+        "id": 21359,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001986.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001986.wav"
+            ]
+        }
+    },
+    {
+        "id": 21360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001908.wav"
+            ]
+        }
+    },
+    {
+        "id": 21365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005139.wav"
+            ]
+        }
+    },
+    {
+        "id": 21374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bagpipe in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004883.wav"
+            ]
+        }
+    },
+    {
+        "id": 21375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005455.wav"
+            ]
+        }
+    },
+    {
+        "id": 21377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008298.wav"
+            ]
+        }
+    },
+    {
+        "id": 21379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000937.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000937.wav"
+            ]
+        }
+    },
+    {
+        "id": 21383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006226.wav"
+            ]
+        }
+    },
+    {
+        "id": 21388,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006663.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006663.wav"
+            ]
+        }
+    },
+    {
+        "id": 21391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002139.wav"
+            ]
+        }
+    },
+    {
+        "id": 21397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004251.wav"
+            ]
+        }
+    },
+    {
+        "id": 21398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008297.wav"
+            ]
+        }
+    },
+    {
+        "id": 21404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003319.wav"
+            ]
+        }
+    },
+    {
+        "id": 21409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002896.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002896.wav"
+            ]
+        }
+    },
+    {
+        "id": 21412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003092.wav"
+            ]
+        }
+    },
+    {
+        "id": 21419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008594.wav"
+            ]
+        }
+    },
+    {
+        "id": 21420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004022.wav"
+            ]
+        }
+    },
+    {
+        "id": 21426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000147.wav"
+            ]
+        }
+    },
+    {
+        "id": 21429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003973.wav"
+            ]
+        }
+    },
+    {
+        "id": 21448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007140.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007140.wav"
+            ]
+        }
+    },
+    {
+        "id": 21450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005204.wav"
+            ]
+        }
+    },
+    {
+        "id": 21453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000027.wav"
+            ]
+        }
+    },
+    {
+        "id": 21458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004991.wav"
+            ]
+        }
+    },
+    {
+        "id": 21464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000258.wav"
+            ]
+        }
+    },
+    {
+        "id": 21471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006815.wav"
+            ]
+        }
+    },
+    {
+        "id": 21475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004087.wav"
+            ]
+        }
+    },
+    {
+        "id": 21476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000936.wav"
+            ]
+        }
+    },
+    {
+        "id": 21478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007028.wav"
+            ]
+        }
+    },
+    {
+        "id": 21481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002489.wav"
+            ]
+        }
+    },
+    {
+        "id": 21482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006024.wav"
+            ]
+        }
+    },
+    {
+        "id": 21485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002549.wav"
+            ]
+        }
+    },
+    {
+        "id": 21487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001605.wav"
+            ]
+        }
+    },
+    {
+        "id": 21499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002710.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002710.wav"
+            ]
+        }
+    },
+    {
+        "id": 21500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005762.wav"
+            ]
+        }
+    },
+    {
+        "id": 21505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006360.wav"
+            ]
+        }
+    },
+    {
+        "id": 21509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a saxophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005007.wav"
+            ]
+        }
+    },
+    {
+        "id": 21521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000070.wav"
+            ]
+        }
+    },
+    {
+        "id": 21525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002182.wav"
+            ]
+        }
+    },
+    {
+        "id": 21526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003797.wav"
+            ]
+        }
+    },
+    {
+        "id": 21527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003549.wav"
+            ]
+        }
+    },
+    {
+        "id": 21531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007949.wav"
+            ]
+        }
+    },
+    {
+        "id": 21534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000107.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000107.wav"
+            ]
+        }
+    },
+    {
+        "id": 21535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008542.wav"
+            ]
+        }
+    },
+    {
+        "id": 21538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002496.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002496.wav"
+            ]
+        }
+    },
+    {
+        "id": 21555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008740.wav"
+            ]
+        }
+    },
+    {
+        "id": 21556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002983.wav"
+            ]
+        }
+    },
+    {
+        "id": 21567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007567.wav"
+            ]
+        }
+    },
+    {
+        "id": 21574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005863.wav"
+            ]
+        }
+    },
+    {
+        "id": 21579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000733.wav"
+            ]
+        }
+    },
+    {
+        "id": 21586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006431.wav"
+            ]
+        }
+    },
+    {
+        "id": 21587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003018.wav"
+            ]
+        }
+    },
+    {
+        "id": 21591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001829.wav"
+            ]
+        }
+    },
+    {
+        "id": 21598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008051.wav"
+            ]
+        }
+    },
+    {
+        "id": 21599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003305.wav"
+            ]
+        }
+    },
+    {
+        "id": 21606,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008559.wav"
+            ]
+        }
+    },
+    {
+        "id": 21608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002781.wav"
+            ]
+        }
+    },
+    {
+        "id": 21614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005966.wav"
+            ]
+        }
+    },
+    {
+        "id": 21617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002982.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002982.wav"
+            ]
+        }
+    },
+    {
+        "id": 21618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007094.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007094.wav"
+            ]
+        }
+    },
+    {
+        "id": 21620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004878.wav"
+            ]
+        }
+    },
+    {
+        "id": 21625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002962.wav"
+            ]
+        }
+    },
+    {
+        "id": 21629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008383.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008383.wav"
+            ]
+        }
+    },
+    {
+        "id": 21633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004684.wav"
+            ]
+        }
+    },
+    {
+        "id": 21635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002455.wav"
+            ]
+        }
+    },
+    {
+        "id": 21639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007538.wav"
+            ]
+        }
+    },
+    {
+        "id": 21640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000265.wav"
+            ]
+        }
+    },
+    {
+        "id": 21641,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003644.wav"
+            ]
+        }
+    },
+    {
+        "id": 21645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005847.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005847.wav"
+            ]
+        }
+    },
+    {
+        "id": 21649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005505.wav"
+            ]
+        }
+    },
+    {
+        "id": 21664,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005605.wav"
+            ]
+        }
+    },
+    {
+        "id": 21677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007479.wav"
+            ]
+        }
+    },
+    {
+        "id": 21682,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007268.wav"
+            ]
+        }
+    },
+    {
+        "id": 21684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007660.wav"
+            ]
+        }
+    },
+    {
+        "id": 21717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006482.wav"
+            ]
+        }
+    },
+    {
+        "id": 21727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003749.wav"
+            ]
+        }
+    },
+    {
+        "id": 21738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001206.wav"
+            ]
+        }
+    },
+    {
+        "id": 21744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002417.wav"
+            ]
+        }
+    },
+    {
+        "id": 21748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008546.wav"
+            ]
+        }
+    },
+    {
+        "id": 21753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007625.wav"
+            ]
+        }
+    },
+    {
+        "id": 21755,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006022.wav"
+            ]
+        }
+    },
+    {
+        "id": 21756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005404.wav"
+            ]
+        }
+    },
+    {
+        "id": 21758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000865.wav"
+            ]
+        }
+    },
+    {
+        "id": 21764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008190.wav"
+            ]
+        }
+    },
+    {
+        "id": 21766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006139.wav"
+            ]
+        }
+    },
+    {
+        "id": 21776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004879.wav"
+            ]
+        }
+    },
+    {
+        "id": 21780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006438.wav"
+            ]
+        }
+    },
+    {
+        "id": 21783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001803.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001803.wav"
+            ]
+        }
+    },
+    {
+        "id": 21787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003395.wav"
+            ]
+        }
+    },
+    {
+        "id": 21802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007634.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007634.wav"
+            ]
+        }
+    },
+    {
+        "id": 21805,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007168.wav"
+            ]
+        }
+    },
+    {
+        "id": 21808,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich acoustic_guitar makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001380.wav"
+            ]
+        }
+    },
+    {
+        "id": 21811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005939.wav"
+            ]
+        }
+    },
+    {
+        "id": 21812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich acoustic_guitar makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008577.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008577.wav"
+            ]
+        }
+    },
+    {
+        "id": 21813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008677.wav"
+            ]
+        }
+    },
+    {
+        "id": 21830,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a suona sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000480.wav"
+            ]
+        }
+    },
+    {
+        "id": 21832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001976.wav"
+            ]
+        }
+    },
+    {
+        "id": 21834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007172.wav"
+            ]
+        }
+    },
+    {
+        "id": 21838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004909.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004909.wav"
+            ]
+        }
+    },
+    {
+        "id": 21845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000055.wav"
+            ]
+        }
+    },
+    {
+        "id": 21847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006898.wav"
+            ]
+        }
+    },
+    {
+        "id": 21853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003304.wav"
+            ]
+        }
+    },
+    {
+        "id": 21855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001084.wav"
+            ]
+        }
+    },
+    {
+        "id": 21857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003616.wav"
+            ]
+        }
+    },
+    {
+        "id": 21859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008799.wav"
+            ]
+        }
+    },
+    {
+        "id": 21860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003635.wav"
+            ]
+        }
+    },
+    {
+        "id": 21861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008619.wav"
+            ]
+        }
+    },
+    {
+        "id": 21882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001981.wav"
+            ]
+        }
+    },
+    {
+        "id": 21885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004165.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004165.wav"
+            ]
+        }
+    },
+    {
+        "id": 21900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003784.wav"
+            ]
+        }
+    },
+    {
+        "id": 21909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002343.wav"
+            ]
+        }
+    },
+    {
+        "id": 21910,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006800.wav"
+            ]
+        }
+    },
+    {
+        "id": 21915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005218.wav"
+            ]
+        }
+    },
+    {
+        "id": 21919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007551.wav"
+            ]
+        }
+    },
+    {
+        "id": 21921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002105.wav"
+            ]
+        }
+    },
+    {
+        "id": 21926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a electric_bass in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001561.wav"
+            ]
+        }
+    },
+    {
+        "id": 21928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007062.wav"
+            ]
+        }
+    },
+    {
+        "id": 21933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008590.wav"
+            ]
+        }
+    },
+    {
+        "id": 21937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006416.wav"
+            ]
+        }
+    },
+    {
+        "id": 21939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003598.wav"
+            ]
+        }
+    },
+    {
+        "id": 21948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007081.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007081.wav"
+            ]
+        }
+    },
+    {
+        "id": 21949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008898.wav"
+            ]
+        }
+    },
+    {
+        "id": 21956,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002382.wav"
+            ]
+        }
+    },
+    {
+        "id": 21961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005875.wav"
+            ]
+        }
+    },
+    {
+        "id": 21964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000960.wav"
+            ]
+        }
+    },
+    {
+        "id": 21973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001816.wav"
+            ]
+        }
+    },
+    {
+        "id": 21984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003171.wav"
+            ]
+        }
+    },
+    {
+        "id": 21988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005219.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005219.wav"
+            ]
+        }
+    },
+    {
+        "id": 21999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000725.wav"
+            ]
+        }
+    },
+    {
+        "id": 22001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005957.wav"
+            ]
+        }
+    },
+    {
+        "id": 22014,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002338.wav"
+            ]
+        }
+    },
+    {
+        "id": 22019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001608.wav"
+            ]
+        }
+    },
+    {
+        "id": 22035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008623.wav"
+            ]
+        }
+    },
+    {
+        "id": 22036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003355.wav"
+            ]
+        }
+    },
+    {
+        "id": 22038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004446.wav"
+            ]
+        }
+    },
+    {
+        "id": 22044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001864.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001864.wav"
+            ]
+        }
+    },
+    {
+        "id": 22058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008965.wav"
+            ]
+        }
+    },
+    {
+        "id": 22062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004918.wav"
+            ]
+        }
+    },
+    {
+        "id": 22071,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002187.wav"
+            ]
+        }
+    },
+    {
+        "id": 22073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000163.wav"
+            ]
+        }
+    },
+    {
+        "id": 22087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000540.wav"
+            ]
+        }
+    },
+    {
+        "id": 22089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007887.wav"
+            ]
+        }
+    },
+    {
+        "id": 22093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a drum in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003777.wav"
+            ]
+        }
+    },
+    {
+        "id": 22096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006365.wav"
+            ]
+        }
+    },
+    {
+        "id": 22098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006135.wav"
+            ]
+        }
+    },
+    {
+        "id": 22108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005792.wav"
+            ]
+        }
+    },
+    {
+        "id": 22114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008973.wav"
+            ]
+        }
+    },
+    {
+        "id": 22116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008935.wav"
+            ]
+        }
+    },
+    {
+        "id": 22123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001599.wav"
+            ]
+        }
+    },
+    {
+        "id": 22125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004014.wav"
+            ]
+        }
+    },
+    {
+        "id": 22127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008655.wav"
+            ]
+        }
+    },
+    {
+        "id": 22128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002227.wav"
+            ]
+        }
+    },
+    {
+        "id": 22132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right more rhythmic than the clarinet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007204.wav"
+            ]
+        }
+    },
+    {
+        "id": 22139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003569.wav"
+            ]
+        }
+    },
+    {
+        "id": 22144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006380.wav"
+            ]
+        }
+    },
+    {
+        "id": 22166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008141.wav"
+            ]
+        }
+    },
+    {
+        "id": 22168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007177.wav"
+            ]
+        }
+    },
+    {
+        "id": 22170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004876.wav"
+            ]
+        }
+    },
+    {
+        "id": 22174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000115.wav"
+            ]
+        }
+    },
+    {
+        "id": 22178,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006226.wav"
+            ]
+        }
+    },
+    {
+        "id": 22180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006415.wav"
+            ]
+        }
+    },
+    {
+        "id": 22202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004362.wav"
+            ]
+        }
+    },
+    {
+        "id": 22209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007448.wav"
+            ]
+        }
+    },
+    {
+        "id": 22211,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008507.wav"
+            ]
+        }
+    },
+    {
+        "id": 22212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004632.wav"
+            ]
+        }
+    },
+    {
+        "id": 22214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004688.wav"
+            ]
+        }
+    },
+    {
+        "id": 22220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007410.wav"
+            ]
+        }
+    },
+    {
+        "id": 22221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003518.wav"
+            ]
+        }
+    },
+    {
+        "id": 22222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006193.wav"
+            ]
+        }
+    },
+    {
+        "id": 22224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008112.wav"
+            ]
+        }
+    },
+    {
+        "id": 22229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001174.wav"
+            ]
+        }
+    },
+    {
+        "id": 22239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004394.wav"
+            ]
+        }
+    },
+    {
+        "id": 22242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007679.wav"
+            ]
+        }
+    },
+    {
+        "id": 22263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007946.wav"
+            ]
+        }
+    },
+    {
+        "id": 22264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003991.wav"
+            ]
+        }
+    },
+    {
+        "id": 22267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002700.wav"
+            ]
+        }
+    },
+    {
+        "id": 22273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003054.wav"
+            ]
+        }
+    },
+    {
+        "id": 22278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004435.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004435.wav"
+            ]
+        }
+    },
+    {
+        "id": 22282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005334.wav"
+            ]
+        }
+    },
+    {
+        "id": 22283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000555.wav"
+            ]
+        }
+    },
+    {
+        "id": 22285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many banjo are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001554.wav"
+            ]
+        }
+    },
+    {
+        "id": 22296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002492.wav"
+            ]
+        }
+    },
+    {
+        "id": 22305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004333.wav"
+            ]
+        }
+    },
+    {
+        "id": 22307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000414.wav"
+            ]
+        }
+    },
+    {
+        "id": 22311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003850.wav"
+            ]
+        }
+    },
+    {
+        "id": 22314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001267.wav"
+            ]
+        }
+    },
+    {
+        "id": 22318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006282.wav"
+            ]
+        }
+    },
+    {
+        "id": 22326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002811.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002811.wav"
+            ]
+        }
+    },
+    {
+        "id": 22335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000921.wav"
+            ]
+        }
+    },
+    {
+        "id": 22337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000815.wav"
+            ]
+        }
+    },
+    {
+        "id": 22348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007299.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007299.wav"
+            ]
+        }
+    },
+    {
+        "id": 22353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004133.wav"
+            ]
+        }
+    },
+    {
+        "id": 22361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006139.wav"
+            ]
+        }
+    },
+    {
+        "id": 22368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007092.wav"
+            ]
+        }
+    },
+    {
+        "id": 22369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004691.wav"
+            ]
+        }
+    },
+    {
+        "id": 22370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many clarinet are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003095.wav"
+            ]
+        }
+    },
+    {
+        "id": 22372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008740.wav"
+            ]
+        }
+    },
+    {
+        "id": 22377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001105.wav"
+            ]
+        }
+    },
+    {
+        "id": 22378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008658.wav"
+            ]
+        }
+    },
+    {
+        "id": 22380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many ukulele are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004844.wav"
+            ]
+        }
+    },
+    {
+        "id": 22381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008858.wav"
+            ]
+        }
+    },
+    {
+        "id": 22382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000246.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000246.wav"
+            ]
+        }
+    },
+    {
+        "id": 22384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001001.wav"
+            ]
+        }
+    },
+    {
+        "id": 22402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008010.wav"
+            ]
+        }
+    },
+    {
+        "id": 22409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003007.wav"
+            ]
+        }
+    },
+    {
+        "id": 22410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004769.wav"
+            ]
+        }
+    },
+    {
+        "id": 22424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006894.wav"
+            ]
+        }
+    },
+    {
+        "id": 22433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001647.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001647.wav"
+            ]
+        }
+    },
+    {
+        "id": 22443,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004337.wav"
+            ]
+        }
+    },
+    {
+        "id": 22448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007418.wav"
+            ]
+        }
+    },
+    {
+        "id": 22451,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many bassoon are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008813.wav"
+            ]
+        }
+    },
+    {
+        "id": 22460,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006599.wav"
+            ]
+        }
+    },
+    {
+        "id": 22463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006094.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006094.wav"
+            ]
+        }
+    },
+    {
+        "id": 22467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001551.wav"
+            ]
+        }
+    },
+    {
+        "id": 22472,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003740.wav"
+            ]
+        }
+    },
+    {
+        "id": 22477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008519.wav"
+            ]
+        }
+    },
+    {
+        "id": 22483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005643.wav"
+            ]
+        }
+    },
+    {
+        "id": 22486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000849.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000849.wav"
+            ]
+        }
+    },
+    {
+        "id": 22489,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003193.wav"
+            ]
+        }
+    },
+    {
+        "id": 22491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008933.wav"
+            ]
+        }
+    },
+    {
+        "id": 22492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004203.wav"
+            ]
+        }
+    },
+    {
+        "id": 22493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004631.wav"
+            ]
+        }
+    },
+    {
+        "id": 22495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004155.wav"
+            ]
+        }
+    },
+    {
+        "id": 22499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the bassoon on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007242.wav"
+            ]
+        }
+    },
+    {
+        "id": 22501,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003381.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003381.wav"
+            ]
+        }
+    },
+    {
+        "id": 22504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006801.wav"
+            ]
+        }
+    },
+    {
+        "id": 22507,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001002.wav"
+            ]
+        }
+    },
+    {
+        "id": 22509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004506.wav"
+            ]
+        }
+    },
+    {
+        "id": 22510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002931.wav"
+            ]
+        }
+    },
+    {
+        "id": 22514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001328.wav"
+            ]
+        }
+    },
+    {
+        "id": 22515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich banjo makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002837.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002837.wav"
+            ]
+        }
+    },
+    {
+        "id": 22521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006334.wav"
+            ]
+        }
+    },
+    {
+        "id": 22523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000461.wav"
+            ]
+        }
+    },
+    {
+        "id": 22524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006807.wav"
+            ]
+        }
+    },
+    {
+        "id": 22534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002532.wav"
+            ]
+        }
+    },
+    {
+        "id": 22538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002586.wav"
+            ]
+        }
+    },
+    {
+        "id": 22540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005328.wav"
+            ]
+        }
+    },
+    {
+        "id": 22542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002163.wav"
+            ]
+        }
+    },
+    {
+        "id": 22552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007365.wav"
+            ]
+        }
+    },
+    {
+        "id": 22558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000040.wav"
+            ]
+        }
+    },
+    {
+        "id": 22559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003800.wav"
+            ]
+        }
+    },
+    {
+        "id": 22574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008029.wav"
+            ]
+        }
+    },
+    {
+        "id": 22585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007425.wav"
+            ]
+        }
+    },
+    {
+        "id": 22594,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000472.wav"
+            ]
+        }
+    },
+    {
+        "id": 22597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003724.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003724.wav"
+            ]
+        }
+    },
+    {
+        "id": 22609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001210.wav"
+            ]
+        }
+    },
+    {
+        "id": 22613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many piano are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002280.wav"
+            ]
+        }
+    },
+    {
+        "id": 22614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001021.wav"
+            ]
+        }
+    },
+    {
+        "id": 22615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008076.wav"
+            ]
+        }
+    },
+    {
+        "id": 22618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001656.wav"
+            ]
+        }
+    },
+    {
+        "id": 22627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a acoustic_guitar in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007757.wav"
+            ]
+        }
+    },
+    {
+        "id": 22631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001041.wav"
+            ]
+        }
+    },
+    {
+        "id": 22638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007895.wav"
+            ]
+        }
+    },
+    {
+        "id": 22640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many cello are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003331.wav"
+            ]
+        }
+    },
+    {
+        "id": 22643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008093.wav"
+            ]
+        }
+    },
+    {
+        "id": 22645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001451.wav"
+            ]
+        }
+    },
+    {
+        "id": 22648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001258.wav"
+            ]
+        }
+    },
+    {
+        "id": 22663,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left louder than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007639.wav"
+            ]
+        }
+    },
+    {
+        "id": 22666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004529.wav"
+            ]
+        }
+    },
+    {
+        "id": 22672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001648.wav"
+            ]
+        }
+    },
+    {
+        "id": 22684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004855.wav"
+            ]
+        }
+    },
+    {
+        "id": 22692,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002659.wav"
+            ]
+        }
+    },
+    {
+        "id": 22695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008223.wav"
+            ]
+        }
+    },
+    {
+        "id": 22699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003578.wav"
+            ]
+        }
+    },
+    {
+        "id": 22708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007506.wav"
+            ]
+        }
+    },
+    {
+        "id": 22728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004951.wav"
+            ]
+        }
+    },
+    {
+        "id": 22734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001150.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001150.wav"
+            ]
+        }
+    },
+    {
+        "id": 22737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007278.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007278.wav"
+            ]
+        }
+    },
+    {
+        "id": 22749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many erhu are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006562.wav"
+            ]
+        }
+    },
+    {
+        "id": 22750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many acoustic_guitar are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004121.wav"
+            ]
+        }
+    },
+    {
+        "id": 22751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004886.wav"
+            ]
+        }
+    },
+    {
+        "id": 22752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006843.wav"
+            ]
+        }
+    },
+    {
+        "id": 22754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002879.wav"
+            ]
+        }
+    },
+    {
+        "id": 22759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007787.wav"
+            ]
+        }
+    },
+    {
+        "id": 22767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007723.wav"
+            ]
+        }
+    },
+    {
+        "id": 22777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007689.wav"
+            ]
+        }
+    },
+    {
+        "id": 22785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002651.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002651.wav"
+            ]
+        }
+    },
+    {
+        "id": 22788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007529.wav"
+            ]
+        }
+    },
+    {
+        "id": 22790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002698.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002698.wav"
+            ]
+        }
+    },
+    {
+        "id": 22800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007840.wav"
+            ]
+        }
+    },
+    {
+        "id": 22802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003364.wav"
+            ]
+        }
+    },
+    {
+        "id": 22803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005513.wav"
+            ]
+        }
+    },
+    {
+        "id": 22804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a flute in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003691.wav"
+            ]
+        }
+    },
+    {
+        "id": 22809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006114.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006114.wav"
+            ]
+        }
+    },
+    {
+        "id": 22810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008065.wav"
+            ]
+        }
+    },
+    {
+        "id": 22814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006767.wav"
+            ]
+        }
+    },
+    {
+        "id": 22815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002722.wav"
+            ]
+        }
+    },
+    {
+        "id": 22816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000562.wav"
+            ]
+        }
+    },
+    {
+        "id": 22820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007218.wav"
+            ]
+        }
+    },
+    {
+        "id": 22823,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001353.wav"
+            ]
+        }
+    },
+    {
+        "id": 22825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003422.wav"
+            ]
+        }
+    },
+    {
+        "id": 22827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000609.wav"
+            ]
+        }
+    },
+    {
+        "id": 22831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007492.wav"
+            ]
+        }
+    },
+    {
+        "id": 22835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004033.wav"
+            ]
+        }
+    },
+    {
+        "id": 22836,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000279.wav"
+            ]
+        }
+    },
+    {
+        "id": 22838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003361.wav"
+            ]
+        }
+    },
+    {
+        "id": 22839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006355.wav"
+            ]
+        }
+    },
+    {
+        "id": 22841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left louder than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008882.wav"
+            ]
+        }
+    },
+    {
+        "id": 22844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004433.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004433.wav"
+            ]
+        }
+    },
+    {
+        "id": 22845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008035.wav"
+            ]
+        }
+    },
+    {
+        "id": 22855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008353.wav"
+            ]
+        }
+    },
+    {
+        "id": 22856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004319.wav"
+            ]
+        }
+    },
+    {
+        "id": 22860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006498.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006498.wav"
+            ]
+        }
+    },
+    {
+        "id": 22864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006120.wav"
+            ]
+        }
+    },
+    {
+        "id": 22867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004286.wav"
+            ]
+        }
+    },
+    {
+        "id": 22872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007259.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007259.wav"
+            ]
+        }
+    },
+    {
+        "id": 22879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007302.wav"
+            ]
+        }
+    },
+    {
+        "id": 22881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many violin are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004567.wav"
+            ]
+        }
+    },
+    {
+        "id": 22885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002436.wav"
+            ]
+        }
+    },
+    {
+        "id": 22886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006019.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006019.wav"
+            ]
+        }
+    },
+    {
+        "id": 22887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002751.wav"
+            ]
+        }
+    },
+    {
+        "id": 22888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007662.wav"
+            ]
+        }
+    },
+    {
+        "id": 22896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007906.wav"
+            ]
+        }
+    },
+    {
+        "id": 22897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many electric_bass are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001546.wav"
+            ]
+        }
+    },
+    {
+        "id": 22898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000077.wav"
+            ]
+        }
+    },
+    {
+        "id": 22902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001729.wav"
+            ]
+        }
+    },
+    {
+        "id": 22903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004501.wav"
+            ]
+        }
+    },
+    {
+        "id": 22904,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001205.wav"
+            ]
+        }
+    },
+    {
+        "id": 22908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003646.wav"
+            ]
+        }
+    },
+    {
+        "id": 22915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003579.wav"
+            ]
+        }
+    },
+    {
+        "id": 22918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003979.wav"
+            ]
+        }
+    },
+    {
+        "id": 22929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002202.wav"
+            ]
+        }
+    },
+    {
+        "id": 22936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006439.wav"
+            ]
+        }
+    },
+    {
+        "id": 22938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002859.wav"
+            ]
+        }
+    },
+    {
+        "id": 22946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006188.wav"
+            ]
+        }
+    },
+    {
+        "id": 22947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002810.wav"
+            ]
+        }
+    },
+    {
+        "id": 22950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001834.wav"
+            ]
+        }
+    },
+    {
+        "id": 22951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006633.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006633.wav"
+            ]
+        }
+    },
+    {
+        "id": 22952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001425.wav"
+            ]
+        }
+    },
+    {
+        "id": 22960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many tuba are in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000113.wav"
+            ]
+        }
+    },
+    {
+        "id": 22961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000162.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000162.wav"
+            ]
+        }
+    },
+    {
+        "id": 22970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004590.wav"
+            ]
+        }
+    },
+    {
+        "id": 22984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006769.wav"
+            ]
+        }
+    },
+    {
+        "id": 22985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000677.wav"
+            ]
+        }
+    },
+    {
+        "id": 22987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001085.wav"
+            ]
+        }
+    },
+    {
+        "id": 22995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007878.wav"
+            ]
+        }
+    },
+    {
+        "id": 23004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002495.wav"
+            ]
+        }
+    },
+    {
+        "id": 23005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006692.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006692.wav"
+            ]
+        }
+    },
+    {
+        "id": 23007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005236.wav"
+            ]
+        }
+    },
+    {
+        "id": 23008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich is the musical instrument that sounds at the same time as the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004675.wav"
+            ]
+        }
+    },
+    {
+        "id": 23012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001453.wav"
+            ]
+        }
+    },
+    {
+        "id": 23016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008102.wav"
+            ]
+        }
+    },
+    {
+        "id": 23025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008003.wav"
+            ]
+        }
+    },
+    {
+        "id": 23030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001126.wav"
+            ]
+        }
+    },
+    {
+        "id": 23032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002517.wav"
+            ]
+        }
+    },
+    {
+        "id": 23042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007188.wav"
+            ]
+        }
+    },
+    {
+        "id": 23048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006300.wav"
+            ]
+        }
+    },
+    {
+        "id": 23054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006120.wav"
+            ]
+        }
+    },
+    {
+        "id": 23064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008297.wav"
+            ]
+        }
+    },
+    {
+        "id": 23078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008437.wav"
+            ]
+        }
+    },
+    {
+        "id": 23086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008892.wav"
+            ]
+        }
+    },
+    {
+        "id": 23088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008525.wav"
+            ]
+        }
+    },
+    {
+        "id": 23093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006799.wav"
+            ]
+        }
+    },
+    {
+        "id": 23095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001490.wav"
+            ]
+        }
+    },
+    {
+        "id": 23099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007652.wav"
+            ]
+        }
+    },
+    {
+        "id": 23101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004585.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004585.wav"
+            ]
+        }
+    },
+    {
+        "id": 23111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001690.wav"
+            ]
+        }
+    },
+    {
+        "id": 23117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004408.wav"
+            ]
+        }
+    },
+    {
+        "id": 23118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008959.wav"
+            ]
+        }
+    },
+    {
+        "id": 23131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003033.wav"
+            ]
+        }
+    },
+    {
+        "id": 23132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005621.wav"
+            ]
+        }
+    },
+    {
+        "id": 23140,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002269.wav"
+            ]
+        }
+    },
+    {
+        "id": 23145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008450.wav"
+            ]
+        }
+    },
+    {
+        "id": 23153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004306.wav"
+            ]
+        }
+    },
+    {
+        "id": 23154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005814.wav"
+            ]
+        }
+    },
+    {
+        "id": 23161,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007365.wav"
+            ]
+        }
+    },
+    {
+        "id": 23166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000709.wav"
+            ]
+        }
+    },
+    {
+        "id": 23171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002260.wav"
+            ]
+        }
+    },
+    {
+        "id": 23203,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003263.wav"
+            ]
+        }
+    },
+    {
+        "id": 23206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001910.wav"
+            ]
+        }
+    },
+    {
+        "id": 23208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and clarinet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003944.wav"
+            ]
+        }
+    },
+    {
+        "id": 23213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000897.wav"
+            ]
+        }
+    },
+    {
+        "id": 23226,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000273.wav"
+            ]
+        }
+    },
+    {
+        "id": 23230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006363.wav"
+            ]
+        }
+    },
+    {
+        "id": 23235,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006161.wav"
+            ]
+        }
+    },
+    {
+        "id": 23239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001561.wav"
+            ]
+        }
+    },
+    {
+        "id": 23242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000244.wav"
+            ]
+        }
+    },
+    {
+        "id": 23254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003154.wav"
+            ]
+        }
+    },
+    {
+        "id": 23261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002179.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002179.wav"
+            ]
+        }
+    },
+    {
+        "id": 23269,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003598.wav"
+            ]
+        }
+    },
+    {
+        "id": 23271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004368.wav"
+            ]
+        }
+    },
+    {
+        "id": 23272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000771.wav"
+            ]
+        }
+    },
+    {
+        "id": 23276,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000749.wav"
+            ]
+        }
+    },
+    {
+        "id": 23278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001857.wav"
+            ]
+        }
+    },
+    {
+        "id": 23280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001980.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001980.wav"
+            ]
+        }
+    },
+    {
+        "id": 23283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001559.wav"
+            ]
+        }
+    },
+    {
+        "id": 23287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003967.wav"
+            ]
+        }
+    },
+    {
+        "id": 23291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005180.wav"
+            ]
+        }
+    },
+    {
+        "id": 23293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005382.wav"
+            ]
+        }
+    },
+    {
+        "id": 23302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000423.wav"
+            ]
+        }
+    },
+    {
+        "id": 23307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005597.wav"
+            ]
+        }
+    },
+    {
+        "id": 23309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008572.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008572.wav"
+            ]
+        }
+    },
+    {
+        "id": 23311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008475.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008475.wav"
+            ]
+        }
+    },
+    {
+        "id": 23313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003031.wav"
+            ]
+        }
+    },
+    {
+        "id": 23316,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000788.wav"
+            ]
+        }
+    },
+    {
+        "id": 23317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001304.wav"
+            ]
+        }
+    },
+    {
+        "id": 23328,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001407.wav"
+            ]
+        }
+    },
+    {
+        "id": 23329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000362.wav"
+            ]
+        }
+    },
+    {
+        "id": 23337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001298.wav"
+            ]
+        }
+    },
+    {
+        "id": 23344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002043.wav"
+            ]
+        }
+    },
+    {
+        "id": 23348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007261.wav"
+            ]
+        }
+    },
+    {
+        "id": 23349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005280.wav"
+            ]
+        }
+    },
+    {
+        "id": 23355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003202.wav"
+            ]
+        }
+    },
+    {
+        "id": 23366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002392.wav"
+            ]
+        }
+    },
+    {
+        "id": 23370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005472.wav"
+            ]
+        }
+    },
+    {
+        "id": 23375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004237.wav"
+            ]
+        }
+    },
+    {
+        "id": 23377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008938.wav"
+            ]
+        }
+    },
+    {
+        "id": 23380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002928.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002928.wav"
+            ]
+        }
+    },
+    {
+        "id": 23390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007634.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007634.wav"
+            ]
+        }
+    },
+    {
+        "id": 23395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008002.wav"
+            ]
+        }
+    },
+    {
+        "id": 23397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008293.wav"
+            ]
+        }
+    },
+    {
+        "id": 23398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005919.wav"
+            ]
+        }
+    },
+    {
+        "id": 23408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004020.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004020.wav"
+            ]
+        }
+    },
+    {
+        "id": 23409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005204.wav"
+            ]
+        }
+    },
+    {
+        "id": 23416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002114.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002114.wav"
+            ]
+        }
+    },
+    {
+        "id": 23431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002279.wav"
+            ]
+        }
+    },
+    {
+        "id": 23434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008961.wav"
+            ]
+        }
+    },
+    {
+        "id": 23436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004963.wav"
+            ]
+        }
+    },
+    {
+        "id": 23442,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001250.wav"
+            ]
+        }
+    },
+    {
+        "id": 23445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008810.wav"
+            ]
+        }
+    },
+    {
+        "id": 23446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006459.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006459.wav"
+            ]
+        }
+    },
+    {
+        "id": 23447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002378.wav"
+            ]
+        }
+    },
+    {
+        "id": 23453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001565.wav"
+            ]
+        }
+    },
+    {
+        "id": 23458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006082.wav"
+            ]
+        }
+    },
+    {
+        "id": 23466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005654.wav"
+            ]
+        }
+    },
+    {
+        "id": 23467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001656.wav"
+            ]
+        }
+    },
+    {
+        "id": 23471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008605.wav"
+            ]
+        }
+    },
+    {
+        "id": 23482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007858.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007858.wav"
+            ]
+        }
+    },
+    {
+        "id": 23483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002410.wav"
+            ]
+        }
+    },
+    {
+        "id": 23497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001815.wav"
+            ]
+        }
+    },
+    {
+        "id": 23500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005644.wav"
+            ]
+        }
+    },
+    {
+        "id": 23504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007165.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007165.wav"
+            ]
+        }
+    },
+    {
+        "id": 23509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004924.wav"
+            ]
+        }
+    },
+    {
+        "id": 23519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the left louder than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001565.wav"
+            ]
+        }
+    },
+    {
+        "id": 23531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002843.wav"
+            ]
+        }
+    },
+    {
+        "id": 23532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006316.wav"
+            ]
+        }
+    },
+    {
+        "id": 23534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005189.wav"
+            ]
+        }
+    },
+    {
+        "id": 23535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007057.wav"
+            ]
+        }
+    },
+    {
+        "id": 23536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001215.wav"
+            ]
+        }
+    },
+    {
+        "id": 23537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the right more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005375.wav"
+            ]
+        }
+    },
+    {
+        "id": 23542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003706.wav"
+            ]
+        }
+    },
+    {
+        "id": 23547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001699.wav"
+            ]
+        }
+    },
+    {
+        "id": 23552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003234.wav"
+            ]
+        }
+    },
+    {
+        "id": 23560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007991.wav"
+            ]
+        }
+    },
+    {
+        "id": 23561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001961.wav"
+            ]
+        }
+    },
+    {
+        "id": 23567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003550.wav"
+            ]
+        }
+    },
+    {
+        "id": 23579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004202.wav"
+            ]
+        }
+    },
+    {
+        "id": 23583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004268.wav"
+            ]
+        }
+    },
+    {
+        "id": 23591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005460.wav"
+            ]
+        }
+    },
+    {
+        "id": 23595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001455.wav"
+            ]
+        }
+    },
+    {
+        "id": 23597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005095.wav"
+            ]
+        }
+    },
+    {
+        "id": 23614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004308.wav"
+            ]
+        }
+    },
+    {
+        "id": 23627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007583.wav"
+            ]
+        }
+    },
+    {
+        "id": 23636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004681.wav"
+            ]
+        }
+    },
+    {
+        "id": 23638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002044.wav"
+            ]
+        }
+    },
+    {
+        "id": 23648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001931.wav"
+            ]
+        }
+    },
+    {
+        "id": 23653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the middle more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008252.wav"
+            ]
+        }
+    },
+    {
+        "id": 23657,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005561.wav"
+            ]
+        }
+    },
+    {
+        "id": 23658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001681.wav"
+            ]
+        }
+    },
+    {
+        "id": 23660,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000713.wav"
+            ]
+        }
+    },
+    {
+        "id": 23665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005821.wav"
+            ]
+        }
+    },
+    {
+        "id": 23670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002430.wav"
+            ]
+        }
+    },
+    {
+        "id": 23676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006954.wav"
+            ]
+        }
+    },
+    {
+        "id": 23688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008747.wav"
+            ]
+        }
+    },
+    {
+        "id": 23698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there suona and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006576.wav"
+            ]
+        }
+    },
+    {
+        "id": 23700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006131.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006131.wav"
+            ]
+        }
+    },
+    {
+        "id": 23702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000902.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000902.wav"
+            ]
+        }
+    },
+    {
+        "id": 23709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000800.wav"
+            ]
+        }
+    },
+    {
+        "id": 23711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003910.wav"
+            ]
+        }
+    },
+    {
+        "id": 23714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002305.wav"
+            ]
+        }
+    },
+    {
+        "id": 23740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006900.wav"
+            ]
+        }
+    },
+    {
+        "id": 23748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001667.wav"
+            ]
+        }
+    },
+    {
+        "id": 23755,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007812.wav"
+            ]
+        }
+    },
+    {
+        "id": 23756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007897.wav"
+            ]
+        }
+    },
+    {
+        "id": 23766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007822.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007822.wav"
+            ]
+        }
+    },
+    {
+        "id": 23774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004756.wav"
+            ]
+        }
+    },
+    {
+        "id": 23783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007341.wav"
+            ]
+        }
+    },
+    {
+        "id": 23790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007380.wav"
+            ]
+        }
+    },
+    {
+        "id": 23798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006263.wav"
+            ]
+        }
+    },
+    {
+        "id": 23803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000019.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000019.wav"
+            ]
+        }
+    },
+    {
+        "id": 23806,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007167.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007167.wav"
+            ]
+        }
+    },
+    {
+        "id": 23828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000471.wav"
+            ]
+        }
+    },
+    {
+        "id": 23829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005687.wav"
+            ]
+        }
+    },
+    {
+        "id": 23833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008965.wav"
+            ]
+        }
+    },
+    {
+        "id": 23834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003394.wav"
+            ]
+        }
+    },
+    {
+        "id": 23854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001537.wav"
+            ]
+        }
+    },
+    {
+        "id": 23856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008806.wav"
+            ]
+        }
+    },
+    {
+        "id": 23864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007583.wav"
+            ]
+        }
+    },
+    {
+        "id": 23865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007597.wav"
+            ]
+        }
+    },
+    {
+        "id": 23867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004528.wav"
+            ]
+        }
+    },
+    {
+        "id": 23870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006011.wav"
+            ]
+        }
+    },
+    {
+        "id": 23874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000311.wav"
+            ]
+        }
+    },
+    {
+        "id": 23881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002542.wav"
+            ]
+        }
+    },
+    {
+        "id": 23883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003602.wav"
+            ]
+        }
+    },
+    {
+        "id": 23884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007639.wav"
+            ]
+        }
+    },
+    {
+        "id": 23887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007957.wav"
+            ]
+        }
+    },
+    {
+        "id": 23888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007702.wav"
+            ]
+        }
+    },
+    {
+        "id": 23892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008003.wav"
+            ]
+        }
+    },
+    {
+        "id": 23898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006714.wav"
+            ]
+        }
+    },
+    {
+        "id": 23900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007010.wav"
+            ]
+        }
+    },
+    {
+        "id": 23906,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000603.wav"
+            ]
+        }
+    },
+    {
+        "id": 23910,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005801.wav"
+            ]
+        }
+    },
+    {
+        "id": 23918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000596.wav"
+            ]
+        }
+    },
+    {
+        "id": 23924,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006466.wav"
+            ]
+        }
+    },
+    {
+        "id": 23926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001514.wav"
+            ]
+        }
+    },
+    {
+        "id": 23930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006888.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006888.wav"
+            ]
+        }
+    },
+    {
+        "id": 23931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006807.wav"
+            ]
+        }
+    },
+    {
+        "id": 23945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000544.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000544.wav"
+            ]
+        }
+    },
+    {
+        "id": 23947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001340.wav"
+            ]
+        }
+    },
+    {
+        "id": 23951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002968.wav"
+            ]
+        }
+    },
+    {
+        "id": 23956,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006991.wav"
+            ]
+        }
+    },
+    {
+        "id": 23958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004862.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004862.wav"
+            ]
+        }
+    },
+    {
+        "id": 23965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006651.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006651.wav"
+            ]
+        }
+    },
+    {
+        "id": 23968,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006827.wav"
+            ]
+        }
+    },
+    {
+        "id": 23974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008115.wav"
+            ]
+        }
+    },
+    {
+        "id": 23983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002901.wav"
+            ]
+        }
+    },
+    {
+        "id": 23986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004088.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004088.wav"
+            ]
+        }
+    },
+    {
+        "id": 23997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003691.wav"
+            ]
+        }
+    },
+    {
+        "id": 24003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004692.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004692.wav"
+            ]
+        }
+    },
+    {
+        "id": 24005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007985.wav"
+            ]
+        }
+    },
+    {
+        "id": 24007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005050.wav"
+            ]
+        }
+    },
+    {
+        "id": 24018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002001.wav"
+            ]
+        }
+    },
+    {
+        "id": 24030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005447.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005447.wav"
+            ]
+        }
+    },
+    {
+        "id": 24035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007545.wav"
+            ]
+        }
+    },
+    {
+        "id": 24039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005638.wav"
+            ]
+        }
+    },
+    {
+        "id": 24040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008208.wav"
+            ]
+        }
+    },
+    {
+        "id": 24041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007281.wav"
+            ]
+        }
+    },
+    {
+        "id": 24047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004636.wav"
+            ]
+        }
+    },
+    {
+        "id": 24053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003238.wav"
+            ]
+        }
+    },
+    {
+        "id": 24061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008895.wav"
+            ]
+        }
+    },
+    {
+        "id": 24065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007738.wav"
+            ]
+        }
+    },
+    {
+        "id": 24067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004012.wav"
+            ]
+        }
+    },
+    {
+        "id": 24073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001525.wav"
+            ]
+        }
+    },
+    {
+        "id": 24075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003818.wav"
+            ]
+        }
+    },
+    {
+        "id": 24092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001719.wav"
+            ]
+        }
+    },
+    {
+        "id": 24094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005387.wav"
+            ]
+        }
+    },
+    {
+        "id": 24095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002848.wav"
+            ]
+        }
+    },
+    {
+        "id": 24103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002876.wav"
+            ]
+        }
+    },
+    {
+        "id": 24104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006573.wav"
+            ]
+        }
+    },
+    {
+        "id": 24106,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006110.wav"
+            ]
+        }
+    },
+    {
+        "id": 24110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000654.wav"
+            ]
+        }
+    },
+    {
+        "id": 24121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002681.wav"
+            ]
+        }
+    },
+    {
+        "id": 24131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002132.wav"
+            ]
+        }
+    },
+    {
+        "id": 24132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008659.wav"
+            ]
+        }
+    },
+    {
+        "id": 24137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007442.wav"
+            ]
+        }
+    },
+    {
+        "id": 24149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007430.wav"
+            ]
+        }
+    },
+    {
+        "id": 24151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000748.wav"
+            ]
+        }
+    },
+    {
+        "id": 24153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000715.wav"
+            ]
+        }
+    },
+    {
+        "id": 24157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and acoustic_guitar sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008917.wav"
+            ]
+        }
+    },
+    {
+        "id": 24162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001141.wav"
+            ]
+        }
+    },
+    {
+        "id": 24188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left louder than the banjo on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007095.wav"
+            ]
+        }
+    },
+    {
+        "id": 24197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003933.wav"
+            ]
+        }
+    },
+    {
+        "id": 24199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007762.wav"
+            ]
+        }
+    },
+    {
+        "id": 24200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001846.wav"
+            ]
+        }
+    },
+    {
+        "id": 24202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007005.wav"
+            ]
+        }
+    },
+    {
+        "id": 24205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008117.wav"
+            ]
+        }
+    },
+    {
+        "id": 24212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003769.wav"
+            ]
+        }
+    },
+    {
+        "id": 24213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001709.wav"
+            ]
+        }
+    },
+    {
+        "id": 24222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left louder than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001302.wav"
+            ]
+        }
+    },
+    {
+        "id": 24224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and xylophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000787.wav"
+            ]
+        }
+    },
+    {
+        "id": 24232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005140.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005140.wav"
+            ]
+        }
+    },
+    {
+        "id": 24245,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005442.wav"
+            ]
+        }
+    },
+    {
+        "id": 24246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007700.wav"
+            ]
+        }
+    },
+    {
+        "id": 24255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004663.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004663.wav"
+            ]
+        }
+    },
+    {
+        "id": 24259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002875.wav"
+            ]
+        }
+    },
+    {
+        "id": 24263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003493.wav"
+            ]
+        }
+    },
+    {
+        "id": 24264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005030.wav"
+            ]
+        }
+    },
+    {
+        "id": 24267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007969.wav"
+            ]
+        }
+    },
+    {
+        "id": 24271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002115.wav"
+            ]
+        }
+    },
+    {
+        "id": 24273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006875.wav"
+            ]
+        }
+    },
+    {
+        "id": 24274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002965.wav"
+            ]
+        }
+    },
+    {
+        "id": 24278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005688.wav"
+            ]
+        }
+    },
+    {
+        "id": 24284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003897.wav"
+            ]
+        }
+    },
+    {
+        "id": 24289,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008093.wav"
+            ]
+        }
+    },
+    {
+        "id": 24298,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003720.wav"
+            ]
+        }
+    },
+    {
+        "id": 24300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006936.wav"
+            ]
+        }
+    },
+    {
+        "id": 24303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002615.wav"
+            ]
+        }
+    },
+    {
+        "id": 24305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001224.wav"
+            ]
+        }
+    },
+    {
+        "id": 24312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007840.wav"
+            ]
+        }
+    },
+    {
+        "id": 24321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003099.wav"
+            ]
+        }
+    },
+    {
+        "id": 24324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008482.wav"
+            ]
+        }
+    },
+    {
+        "id": 24327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002312.wav"
+            ]
+        }
+    },
+    {
+        "id": 24328,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001174.wav"
+            ]
+        }
+    },
+    {
+        "id": 24331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and bassoon sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002720.wav"
+            ]
+        }
+    },
+    {
+        "id": 24335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005499.wav"
+            ]
+        }
+    },
+    {
+        "id": 24336,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000635.wav"
+            ]
+        }
+    },
+    {
+        "id": 24345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000104.wav"
+            ]
+        }
+    },
+    {
+        "id": 24346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000578.wav"
+            ]
+        }
+    },
+    {
+        "id": 24347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007206.wav"
+            ]
+        }
+    },
+    {
+        "id": 24351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001900.wav"
+            ]
+        }
+    },
+    {
+        "id": 24356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004873.wav"
+            ]
+        }
+    },
+    {
+        "id": 24361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007403.wav"
+            ]
+        }
+    },
+    {
+        "id": 24362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007690.wav"
+            ]
+        }
+    },
+    {
+        "id": 24363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007021.wav"
+            ]
+        }
+    },
+    {
+        "id": 24366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001089.wav"
+            ]
+        }
+    },
+    {
+        "id": 24369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008921.wav"
+            ]
+        }
+    },
+    {
+        "id": 24372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003149.wav"
+            ]
+        }
+    },
+    {
+        "id": 24376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008935.wav"
+            ]
+        }
+    },
+    {
+        "id": 24383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002180.wav"
+            ]
+        }
+    },
+    {
+        "id": 24392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003310.wav"
+            ]
+        }
+    },
+    {
+        "id": 24393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003801.wav"
+            ]
+        }
+    },
+    {
+        "id": 24402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004382.wav"
+            ]
+        }
+    },
+    {
+        "id": 24413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007057.wav"
+            ]
+        }
+    },
+    {
+        "id": 24414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004006.wav"
+            ]
+        }
+    },
+    {
+        "id": 24427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000687.wav"
+            ]
+        }
+    },
+    {
+        "id": 24431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005645.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005645.wav"
+            ]
+        }
+    },
+    {
+        "id": 24432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004031.wav"
+            ]
+        }
+    },
+    {
+        "id": 24434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000801.wav"
+            ]
+        }
+    },
+    {
+        "id": 24437,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004321.wav"
+            ]
+        }
+    },
+    {
+        "id": 24441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001009.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001009.wav"
+            ]
+        }
+    },
+    {
+        "id": 24446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005880.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005880.wav"
+            ]
+        }
+    },
+    {
+        "id": 24448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001120.wav"
+            ]
+        }
+    },
+    {
+        "id": 24461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006365.wav"
+            ]
+        }
+    },
+    {
+        "id": 24470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006527.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006527.wav"
+            ]
+        }
+    },
+    {
+        "id": 24473,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001153.wav"
+            ]
+        }
+    },
+    {
+        "id": 24479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004504.wav"
+            ]
+        }
+    },
+    {
+        "id": 24486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004741.wav"
+            ]
+        }
+    },
+    {
+        "id": 24487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005164.wav"
+            ]
+        }
+    },
+    {
+        "id": 24495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003144.wav"
+            ]
+        }
+    },
+    {
+        "id": 24496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000793.wav"
+            ]
+        }
+    },
+    {
+        "id": 24498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004448.wav"
+            ]
+        }
+    },
+    {
+        "id": 24499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003725.wav"
+            ]
+        }
+    },
+    {
+        "id": 24500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000607.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000607.wav"
+            ]
+        }
+    },
+    {
+        "id": 24503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008516.wav"
+            ]
+        }
+    },
+    {
+        "id": 24515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005197.wav"
+            ]
+        }
+    },
+    {
+        "id": 24522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004610.wav"
+            ]
+        }
+    },
+    {
+        "id": 24526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003414.wav"
+            ]
+        }
+    },
+    {
+        "id": 24529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich congas makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002181.wav"
+            ]
+        }
+    },
+    {
+        "id": 24535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006447.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006447.wav"
+            ]
+        }
+    },
+    {
+        "id": 24536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001722.wav"
+            ]
+        }
+    },
+    {
+        "id": 24540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000973.wav"
+            ]
+        }
+    },
+    {
+        "id": 24543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001718.wav"
+            ]
+        }
+    },
+    {
+        "id": 24545,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003402.wav"
+            ]
+        }
+    },
+    {
+        "id": 24546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003921.wav"
+            ]
+        }
+    },
+    {
+        "id": 24547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007719.wav"
+            ]
+        }
+    },
+    {
+        "id": 24548,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007249.wav"
+            ]
+        }
+    },
+    {
+        "id": 24552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001540.wav"
+            ]
+        }
+    },
+    {
+        "id": 24561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006332.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006332.wav"
+            ]
+        }
+    },
+    {
+        "id": 24568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there suona and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003212.wav"
+            ]
+        }
+    },
+    {
+        "id": 24570,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003484.wav"
+            ]
+        }
+    },
+    {
+        "id": 24571,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007905.wav"
+            ]
+        }
+    },
+    {
+        "id": 24581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000687.wav"
+            ]
+        }
+    },
+    {
+        "id": 24583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001430.wav"
+            ]
+        }
+    },
+    {
+        "id": 24588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006316.wav"
+            ]
+        }
+    },
+    {
+        "id": 24591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right louder than the drum on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005126.wav"
+            ]
+        }
+    },
+    {
+        "id": 24594,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000974.wav"
+            ]
+        }
+    },
+    {
+        "id": 24596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and pipa sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005736.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005736.wav"
+            ]
+        }
+    },
+    {
+        "id": 24600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000718.wav"
+            ]
+        }
+    },
+    {
+        "id": 24608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003722.wav"
+            ]
+        }
+    },
+    {
+        "id": 24609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005732.wav"
+            ]
+        }
+    },
+    {
+        "id": 24614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000603.wav"
+            ]
+        }
+    },
+    {
+        "id": 24624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000833.wav"
+            ]
+        }
+    },
+    {
+        "id": 24626,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003282.wav"
+            ]
+        }
+    },
+    {
+        "id": 24629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005800.wav"
+            ]
+        }
+    },
+    {
+        "id": 24630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004153.wav"
+            ]
+        }
+    },
+    {
+        "id": 24632,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008632.wav"
+            ]
+        }
+    },
+    {
+        "id": 24633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004729.wav"
+            ]
+        }
+    },
+    {
+        "id": 24638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006649.wav"
+            ]
+        }
+    },
+    {
+        "id": 24650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006404.wav"
+            ]
+        }
+    },
+    {
+        "id": 24651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002006.wav"
+            ]
+        }
+    },
+    {
+        "id": 24673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005502.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005502.wav"
+            ]
+        }
+    },
+    {
+        "id": 24674,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003013.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003013.wav"
+            ]
+        }
+    },
+    {
+        "id": 24677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004072.wav"
+            ]
+        }
+    },
+    {
+        "id": 24681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001940.wav"
+            ]
+        }
+    },
+    {
+        "id": 24688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005340.wav"
+            ]
+        }
+    },
+    {
+        "id": 24694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004844.wav"
+            ]
+        }
+    },
+    {
+        "id": 24697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003403.wav"
+            ]
+        }
+    },
+    {
+        "id": 24698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006872.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006872.wav"
+            ]
+        }
+    },
+    {
+        "id": 24699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004428.wav"
+            ]
+        }
+    },
+    {
+        "id": 24702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005939.wav"
+            ]
+        }
+    },
+    {
+        "id": 24703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006402.wav"
+            ]
+        }
+    },
+    {
+        "id": 24704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005883.wav"
+            ]
+        }
+    },
+    {
+        "id": 24713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005608.wav"
+            ]
+        }
+    },
+    {
+        "id": 24718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005015.wav"
+            ]
+        }
+    },
+    {
+        "id": 24722,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005792.wav"
+            ]
+        }
+    },
+    {
+        "id": 24725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the accordion on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007557.wav"
+            ]
+        }
+    },
+    {
+        "id": 24726,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000135.wav"
+            ]
+        }
+    },
+    {
+        "id": 24736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003790.wav"
+            ]
+        }
+    },
+    {
+        "id": 24745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008487.wav"
+            ]
+        }
+    },
+    {
+        "id": 24751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004967.wav"
+            ]
+        }
+    },
+    {
+        "id": 24760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008065.wav"
+            ]
+        }
+    },
+    {
+        "id": 24764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007827.wav"
+            ]
+        }
+    },
+    {
+        "id": 24766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002817.wav"
+            ]
+        }
+    },
+    {
+        "id": 24770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the clarinet on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006396.wav"
+            ]
+        }
+    },
+    {
+        "id": 24771,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004418.wav"
+            ]
+        }
+    },
+    {
+        "id": 24772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004340.wav"
+            ]
+        }
+    },
+    {
+        "id": 24777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005036.wav"
+            ]
+        }
+    },
+    {
+        "id": 24778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000206.wav"
+            ]
+        }
+    },
+    {
+        "id": 24785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002922.wav"
+            ]
+        }
+    },
+    {
+        "id": 24793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004788.wav"
+            ]
+        }
+    },
+    {
+        "id": 24797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000006.wav"
+            ]
+        }
+    },
+    {
+        "id": 24807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004013.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004013.wav"
+            ]
+        }
+    },
+    {
+        "id": 24808,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and drum sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004492.wav"
+            ]
+        }
+    },
+    {
+        "id": 24822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008140.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008140.wav"
+            ]
+        }
+    },
+    {
+        "id": 24831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004605.wav"
+            ]
+        }
+    },
+    {
+        "id": 24832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008060.wav"
+            ]
+        }
+    },
+    {
+        "id": 24834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and congas sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001955.wav"
+            ]
+        }
+    },
+    {
+        "id": 24835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007260.wav"
+            ]
+        }
+    },
+    {
+        "id": 24839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001343.wav"
+            ]
+        }
+    },
+    {
+        "id": 24843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008308.wav"
+            ]
+        }
+    },
+    {
+        "id": 24844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006801.wav"
+            ]
+        }
+    },
+    {
+        "id": 24850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006771.wav"
+            ]
+        }
+    },
+    {
+        "id": 24857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006400.wav"
+            ]
+        }
+    },
+    {
+        "id": 24861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000080.wav"
+            ]
+        }
+    },
+    {
+        "id": 24866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich congas makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008271.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008271.wav"
+            ]
+        }
+    },
+    {
+        "id": 24869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001260.wav"
+            ]
+        }
+    },
+    {
+        "id": 24873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007777.wav"
+            ]
+        }
+    },
+    {
+        "id": 24877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005490.wav"
+            ]
+        }
+    },
+    {
+        "id": 24879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004660.wav"
+            ]
+        }
+    },
+    {
+        "id": 24891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000935.wav"
+            ]
+        }
+    },
+    {
+        "id": 24900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005073.wav"
+            ]
+        }
+    },
+    {
+        "id": 24903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006970.wav"
+            ]
+        }
+    },
+    {
+        "id": 24906,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004045.wav"
+            ]
+        }
+    },
+    {
+        "id": 24908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001267.wav"
+            ]
+        }
+    },
+    {
+        "id": 24914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005109.wav"
+            ]
+        }
+    },
+    {
+        "id": 24918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007244.wav"
+            ]
+        }
+    },
+    {
+        "id": 24921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004763.wav"
+            ]
+        }
+    },
+    {
+        "id": 24926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008358.wav"
+            ]
+        }
+    },
+    {
+        "id": 24930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007307.wav"
+            ]
+        }
+    },
+    {
+        "id": 24937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006715.wav"
+            ]
+        }
+    },
+    {
+        "id": 24940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004358.wav"
+            ]
+        }
+    },
+    {
+        "id": 24951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007725.wav"
+            ]
+        }
+    },
+    {
+        "id": 24966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005890.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005890.wav"
+            ]
+        }
+    },
+    {
+        "id": 24972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001361.wav"
+            ]
+        }
+    },
+    {
+        "id": 24974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001932.wav"
+            ]
+        }
+    },
+    {
+        "id": 24978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005384.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005384.wav"
+            ]
+        }
+    },
+    {
+        "id": 24979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002979.wav"
+            ]
+        }
+    },
+    {
+        "id": 24987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000969.wav"
+            ]
+        }
+    },
+    {
+        "id": 24993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000996.wav"
+            ]
+        }
+    },
+    {
+        "id": 24996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008744.wav"
+            ]
+        }
+    },
+    {
+        "id": 24999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000549.wav"
+            ]
+        }
+    },
+    {
+        "id": 25000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008457.wav"
+            ]
+        }
+    },
+    {
+        "id": 25005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000480.wav"
+            ]
+        }
+    },
+    {
+        "id": 25010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003403.wav"
+            ]
+        }
+    },
+    {
+        "id": 25015,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002188.wav"
+            ]
+        }
+    },
+    {
+        "id": 25016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006974.wav"
+            ]
+        }
+    },
+    {
+        "id": 25019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007604.wav"
+            ]
+        }
+    },
+    {
+        "id": 25023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008106.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008106.wav"
+            ]
+        }
+    },
+    {
+        "id": 25027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001819.wav"
+            ]
+        }
+    },
+    {
+        "id": 25028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000820.wav"
+            ]
+        }
+    },
+    {
+        "id": 25029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003772.wav"
+            ]
+        }
+    },
+    {
+        "id": 25043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001495.wav"
+            ]
+        }
+    },
+    {
+        "id": 25046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008346.wav"
+            ]
+        }
+    },
+    {
+        "id": 25054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003067.wav"
+            ]
+        }
+    },
+    {
+        "id": 25055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003484.wav"
+            ]
+        }
+    },
+    {
+        "id": 25067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002799.wav"
+            ]
+        }
+    },
+    {
+        "id": 25069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007514.wav"
+            ]
+        }
+    },
+    {
+        "id": 25070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001040.wav"
+            ]
+        }
+    },
+    {
+        "id": 25076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005957.wav"
+            ]
+        }
+    },
+    {
+        "id": 25078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008407.wav"
+            ]
+        }
+    },
+    {
+        "id": 25079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008806.wav"
+            ]
+        }
+    },
+    {
+        "id": 25090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001801.wav"
+            ]
+        }
+    },
+    {
+        "id": 25094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left louder than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006692.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006692.wav"
+            ]
+        }
+    },
+    {
+        "id": 25102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000030.wav"
+            ]
+        }
+    },
+    {
+        "id": 25108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008297.wav"
+            ]
+        }
+    },
+    {
+        "id": 25112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and guzheng sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008036.wav"
+            ]
+        }
+    },
+    {
+        "id": 25114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the middle louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006866.wav"
+            ]
+        }
+    },
+    {
+        "id": 25118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003604.wav"
+            ]
+        }
+    },
+    {
+        "id": 25127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003981.wav"
+            ]
+        }
+    },
+    {
+        "id": 25140,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005597.wav"
+            ]
+        }
+    },
+    {
+        "id": 25146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007011.wav"
+            ]
+        }
+    },
+    {
+        "id": 25154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003557.wav"
+            ]
+        }
+    },
+    {
+        "id": 25160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004485.wav"
+            ]
+        }
+    },
+    {
+        "id": 25172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000185.wav"
+            ]
+        }
+    },
+    {
+        "id": 25199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007493.wav"
+            ]
+        }
+    },
+    {
+        "id": 25204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007480.wav"
+            ]
+        }
+    },
+    {
+        "id": 25212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006901.wav"
+            ]
+        }
+    },
+    {
+        "id": 25213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005252.wav"
+            ]
+        }
+    },
+    {
+        "id": 25221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005078.wav"
+            ]
+        }
+    },
+    {
+        "id": 25226,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003555.wav"
+            ]
+        }
+    },
+    {
+        "id": 25231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001444.wav"
+            ]
+        }
+    },
+    {
+        "id": 25235,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004261.wav"
+            ]
+        }
+    },
+    {
+        "id": 25236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003917.wav"
+            ]
+        }
+    },
+    {
+        "id": 25240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005826.wav"
+            ]
+        }
+    },
+    {
+        "id": 25244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000442.wav"
+            ]
+        }
+    },
+    {
+        "id": 25246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005425.wav"
+            ]
+        }
+    },
+    {
+        "id": 25247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002862.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002862.wav"
+            ]
+        }
+    },
+    {
+        "id": 25253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001054.wav"
+            ]
+        }
+    },
+    {
+        "id": 25257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006860.wav"
+            ]
+        }
+    },
+    {
+        "id": 25260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001630.wav"
+            ]
+        }
+    },
+    {
+        "id": 25268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008962.wav"
+            ]
+        }
+    },
+    {
+        "id": 25279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006700.wav"
+            ]
+        }
+    },
+    {
+        "id": 25286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000379.wav"
+            ]
+        }
+    },
+    {
+        "id": 25310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003051.wav"
+            ]
+        }
+    },
+    {
+        "id": 25313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000718.wav"
+            ]
+        }
+    },
+    {
+        "id": 25314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007006.wav"
+            ]
+        }
+    },
+    {
+        "id": 25319,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000356.wav"
+            ]
+        }
+    },
+    {
+        "id": 25327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008503.wav"
+            ]
+        }
+    },
+    {
+        "id": 25337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the right louder than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006787.wav"
+            ]
+        }
+    },
+    {
+        "id": 25345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002117.wav"
+            ]
+        }
+    },
+    {
+        "id": 25347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006885.wav"
+            ]
+        }
+    },
+    {
+        "id": 25349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003647.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003647.wav"
+            ]
+        }
+    },
+    {
+        "id": 25373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006183.wav"
+            ]
+        }
+    },
+    {
+        "id": 25375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001307.wav"
+            ]
+        }
+    },
+    {
+        "id": 25376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005411.wav"
+            ]
+        }
+    },
+    {
+        "id": 25377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002523.wav"
+            ]
+        }
+    },
+    {
+        "id": 25385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005669.wav"
+            ]
+        }
+    },
+    {
+        "id": 25389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004336.wav"
+            ]
+        }
+    },
+    {
+        "id": 25391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004441.wav"
+            ]
+        }
+    },
+    {
+        "id": 25394,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005463.wav"
+            ]
+        }
+    },
+    {
+        "id": 25396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001632.wav"
+            ]
+        }
+    },
+    {
+        "id": 25399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007911.wav"
+            ]
+        }
+    },
+    {
+        "id": 25403,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002461.wav"
+            ]
+        }
+    },
+    {
+        "id": 25413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006115.wav"
+            ]
+        }
+    },
+    {
+        "id": 25415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003195.wav"
+            ]
+        }
+    },
+    {
+        "id": 25417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003208.wav"
+            ]
+        }
+    },
+    {
+        "id": 25418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004881.wav"
+            ]
+        }
+    },
+    {
+        "id": 25423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007600.wav"
+            ]
+        }
+    },
+    {
+        "id": 25428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007988.wav"
+            ]
+        }
+    },
+    {
+        "id": 25433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006637.wav"
+            ]
+        }
+    },
+    {
+        "id": 25437,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006850.wav"
+            ]
+        }
+    },
+    {
+        "id": 25441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005104.wav"
+            ]
+        }
+    },
+    {
+        "id": 25445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001870.wav"
+            ]
+        }
+    },
+    {
+        "id": 25449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005010.wav"
+            ]
+        }
+    },
+    {
+        "id": 25452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004983.wav"
+            ]
+        }
+    },
+    {
+        "id": 25456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006262.wav"
+            ]
+        }
+    },
+    {
+        "id": 25462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008334.wav"
+            ]
+        }
+    },
+    {
+        "id": 25471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007836.wav"
+            ]
+        }
+    },
+    {
+        "id": 25477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007618.wav"
+            ]
+        }
+    },
+    {
+        "id": 25482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007093.wav"
+            ]
+        }
+    },
+    {
+        "id": 25487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004769.wav"
+            ]
+        }
+    },
+    {
+        "id": 25503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and flute sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000006.wav"
+            ]
+        }
+    },
+    {
+        "id": 25509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000921.wav"
+            ]
+        }
+    },
+    {
+        "id": 25513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006607.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006607.wav"
+            ]
+        }
+    },
+    {
+        "id": 25514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008267.wav"
+            ]
+        }
+    },
+    {
+        "id": 25529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006069.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006069.wav"
+            ]
+        }
+    },
+    {
+        "id": 25536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008644.wav"
+            ]
+        }
+    },
+    {
+        "id": 25539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004393.wav"
+            ]
+        }
+    },
+    {
+        "id": 25542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004891.wav"
+            ]
+        }
+    },
+    {
+        "id": 25544,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003045.wav"
+            ]
+        }
+    },
+    {
+        "id": 25554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008125.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008125.wav"
+            ]
+        }
+    },
+    {
+        "id": 25561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001851.wav"
+            ]
+        }
+    },
+    {
+        "id": 25573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and saxophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007557.wav"
+            ]
+        }
+    },
+    {
+        "id": 25578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003191.wav"
+            ]
+        }
+    },
+    {
+        "id": 25580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002712.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002712.wav"
+            ]
+        }
+    },
+    {
+        "id": 25598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003519.wav"
+            ]
+        }
+    },
+    {
+        "id": 25600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008716.wav"
+            ]
+        }
+    },
+    {
+        "id": 25601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001679.wav"
+            ]
+        }
+    },
+    {
+        "id": 25602,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003937.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003937.wav"
+            ]
+        }
+    },
+    {
+        "id": 25607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005283.wav"
+            ]
+        }
+    },
+    {
+        "id": 25609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003790.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003790.wav"
+            ]
+        }
+    },
+    {
+        "id": 25613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004276.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004276.wav"
+            ]
+        }
+    },
+    {
+        "id": 25626,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000845.wav"
+            ]
+        }
+    },
+    {
+        "id": 25633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005724.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005724.wav"
+            ]
+        }
+    },
+    {
+        "id": 25645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002735.wav"
+            ]
+        }
+    },
+    {
+        "id": 25656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003487.wav"
+            ]
+        }
+    },
+    {
+        "id": 25662,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004409.wav"
+            ]
+        }
+    },
+    {
+        "id": 25670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004285.wav"
+            ]
+        }
+    },
+    {
+        "id": 25673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007902.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007902.wav"
+            ]
+        }
+    },
+    {
+        "id": 25676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and congas instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000410.wav"
+            ]
+        }
+    },
+    {
+        "id": 25684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and xylophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007304.wav"
+            ]
+        }
+    },
+    {
+        "id": 25695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007784.wav"
+            ]
+        }
+    },
+    {
+        "id": 25697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005035.wav"
+            ]
+        }
+    },
+    {
+        "id": 25702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004972.wav"
+            ]
+        }
+    },
+    {
+        "id": 25706,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003629.wav"
+            ]
+        }
+    },
+    {
+        "id": 25716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008539.wav"
+            ]
+        }
+    },
+    {
+        "id": 25725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008001.wav"
+            ]
+        }
+    },
+    {
+        "id": 25729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000184.wav"
+            ]
+        }
+    },
+    {
+        "id": 25735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001366.wav"
+            ]
+        }
+    },
+    {
+        "id": 25737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002006.wav"
+            ]
+        }
+    },
+    {
+        "id": 25739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004241.wav"
+            ]
+        }
+    },
+    {
+        "id": 25740,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003624.wav"
+            ]
+        }
+    },
+    {
+        "id": 25745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000996.wav"
+            ]
+        }
+    },
+    {
+        "id": 25746,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006990.wav"
+            ]
+        }
+    },
+    {
+        "id": 25749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006056.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006056.wav"
+            ]
+        }
+    },
+    {
+        "id": 25756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004912.wav"
+            ]
+        }
+    },
+    {
+        "id": 25784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008576.wav"
+            ]
+        }
+    },
+    {
+        "id": 25789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004270.wav"
+            ]
+        }
+    },
+    {
+        "id": 25795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002963.wav"
+            ]
+        }
+    },
+    {
+        "id": 25798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000338.wav"
+            ]
+        }
+    },
+    {
+        "id": 25803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008774.wav"
+            ]
+        }
+    },
+    {
+        "id": 25816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004613.wav"
+            ]
+        }
+    },
+    {
+        "id": 25820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006386.wav"
+            ]
+        }
+    },
+    {
+        "id": 25826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005338.wav"
+            ]
+        }
+    },
+    {
+        "id": 25834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001542.wav"
+            ]
+        }
+    },
+    {
+        "id": 25843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006594.wav"
+            ]
+        }
+    },
+    {
+        "id": 25863,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000593.wav"
+            ]
+        }
+    },
+    {
+        "id": 25868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008045.wav"
+            ]
+        }
+    },
+    {
+        "id": 25871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005999.wav"
+            ]
+        }
+    },
+    {
+        "id": 25878,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008972.wav"
+            ]
+        }
+    },
+    {
+        "id": 25886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007052.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007052.wav"
+            ]
+        }
+    },
+    {
+        "id": 25891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004022.wav"
+            ]
+        }
+    },
+    {
+        "id": 25898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000342.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000342.wav"
+            ]
+        }
+    },
+    {
+        "id": 25903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007843.wav"
+            ]
+        }
+    },
+    {
+        "id": 25904,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007222.wav"
+            ]
+        }
+    },
+    {
+        "id": 25913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004096.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004096.wav"
+            ]
+        }
+    },
+    {
+        "id": 25914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008235.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008235.wav"
+            ]
+        }
+    },
+    {
+        "id": 25918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001781.wav"
+            ]
+        }
+    },
+    {
+        "id": 25927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004060.wav"
+            ]
+        }
+    },
+    {
+        "id": 25936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007926.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007926.wav"
+            ]
+        }
+    },
+    {
+        "id": 25941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003104.wav"
+            ]
+        }
+    },
+    {
+        "id": 25947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000979.wav"
+            ]
+        }
+    },
+    {
+        "id": 25949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004407.wav"
+            ]
+        }
+    },
+    {
+        "id": 25962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002534.wav"
+            ]
+        }
+    },
+    {
+        "id": 25966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001692.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001692.wav"
+            ]
+        }
+    },
+    {
+        "id": 25977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002718.wav"
+            ]
+        }
+    },
+    {
+        "id": 25983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006998.wav"
+            ]
+        }
+    },
+    {
+        "id": 25985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000462.wav"
+            ]
+        }
+    },
+    {
+        "id": 25986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008578.wav"
+            ]
+        }
+    },
+    {
+        "id": 25987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007872.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007872.wav"
+            ]
+        }
+    },
+    {
+        "id": 25997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001713.wav"
+            ]
+        }
+    },
+    {
+        "id": 26004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003160.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003160.wav"
+            ]
+        }
+    },
+    {
+        "id": 26007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008465.wav"
+            ]
+        }
+    },
+    {
+        "id": 26018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006078.wav"
+            ]
+        }
+    },
+    {
+        "id": 26020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006375.wav"
+            ]
+        }
+    },
+    {
+        "id": 26026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002591.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002591.wav"
+            ]
+        }
+    },
+    {
+        "id": 26029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003538.wav"
+            ]
+        }
+    },
+    {
+        "id": 26035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001480.wav"
+            ]
+        }
+    },
+    {
+        "id": 26036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and pipa instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008430.wav"
+            ]
+        }
+    },
+    {
+        "id": 26040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006786.wav"
+            ]
+        }
+    },
+    {
+        "id": 26047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003327.wav"
+            ]
+        }
+    },
+    {
+        "id": 26051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000893.wav"
+            ]
+        }
+    },
+    {
+        "id": 26053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006399.wav"
+            ]
+        }
+    },
+    {
+        "id": 26055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left louder than the accordion on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005730.wav"
+            ]
+        }
+    },
+    {
+        "id": 26065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003352.wav"
+            ]
+        }
+    },
+    {
+        "id": 26067,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and electric_bass sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002288.wav"
+            ]
+        }
+    },
+    {
+        "id": 26072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002365.wav"
+            ]
+        }
+    },
+    {
+        "id": 26079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000346.wav"
+            ]
+        }
+    },
+    {
+        "id": 26089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006885.wav"
+            ]
+        }
+    },
+    {
+        "id": 26091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004310.wav"
+            ]
+        }
+    },
+    {
+        "id": 26100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001402.wav"
+            ]
+        }
+    },
+    {
+        "id": 26107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002448.wav"
+            ]
+        }
+    },
+    {
+        "id": 26108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000200.wav"
+            ]
+        }
+    },
+    {
+        "id": 26117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004671.wav"
+            ]
+        }
+    },
+    {
+        "id": 26124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001275.wav"
+            ]
+        }
+    },
+    {
+        "id": 26126,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001389.wav"
+            ]
+        }
+    },
+    {
+        "id": 26129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000325.wav"
+            ]
+        }
+    },
+    {
+        "id": 26130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005144.wav"
+            ]
+        }
+    },
+    {
+        "id": 26132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002104.wav"
+            ]
+        }
+    },
+    {
+        "id": 26139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006217.wav"
+            ]
+        }
+    },
+    {
+        "id": 26145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005239.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005239.wav"
+            ]
+        }
+    },
+    {
+        "id": 26146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004417.wav"
+            ]
+        }
+    },
+    {
+        "id": 26150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000642.wav"
+            ]
+        }
+    },
+    {
+        "id": 26157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007551.wav"
+            ]
+        }
+    },
+    {
+        "id": 26167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006329.wav"
+            ]
+        }
+    },
+    {
+        "id": 26170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001319.wav"
+            ]
+        }
+    },
+    {
+        "id": 26176,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003287.wav"
+            ]
+        }
+    },
+    {
+        "id": 26187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000652.wav"
+            ]
+        }
+    },
+    {
+        "id": 26215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002030.wav"
+            ]
+        }
+    },
+    {
+        "id": 26219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001837.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001837.wav"
+            ]
+        }
+    },
+    {
+        "id": 26220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005951.wav"
+            ]
+        }
+    },
+    {
+        "id": 26229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000696.wav"
+            ]
+        }
+    },
+    {
+        "id": 26236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008266.wav"
+            ]
+        }
+    },
+    {
+        "id": 26237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005471.wav"
+            ]
+        }
+    },
+    {
+        "id": 26239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003984.wav"
+            ]
+        }
+    },
+    {
+        "id": 26248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006057.wav"
+            ]
+        }
+    },
+    {
+        "id": 26252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006051.wav"
+            ]
+        }
+    },
+    {
+        "id": 26255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002255.wav"
+            ]
+        }
+    },
+    {
+        "id": 26261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005722.wav"
+            ]
+        }
+    },
+    {
+        "id": 26265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003551.wav"
+            ]
+        }
+    },
+    {
+        "id": 26266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008929.wav"
+            ]
+        }
+    },
+    {
+        "id": 26279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007400.wav"
+            ]
+        }
+    },
+    {
+        "id": 26281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006991.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006991.wav"
+            ]
+        }
+    },
+    {
+        "id": 26282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003589.wav"
+            ]
+        }
+    },
+    {
+        "id": 26291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002893.wav"
+            ]
+        }
+    },
+    {
+        "id": 26300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003277.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003277.wav"
+            ]
+        }
+    },
+    {
+        "id": 26308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004599.wav"
+            ]
+        }
+    },
+    {
+        "id": 26309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008241.wav"
+            ]
+        }
+    },
+    {
+        "id": 26314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005911.wav"
+            ]
+        }
+    },
+    {
+        "id": 26323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003993.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003993.wav"
+            ]
+        }
+    },
+    {
+        "id": 26324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007072.wav"
+            ]
+        }
+    },
+    {
+        "id": 26325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007945.wav"
+            ]
+        }
+    },
+    {
+        "id": 26330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008279.wav"
+            ]
+        }
+    },
+    {
+        "id": 26331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003026.wav"
+            ]
+        }
+    },
+    {
+        "id": 26332,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008523.wav"
+            ]
+        }
+    },
+    {
+        "id": 26335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002347.wav"
+            ]
+        }
+    },
+    {
+        "id": 26337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002240.wav"
+            ]
+        }
+    },
+    {
+        "id": 26338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005048.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005048.wav"
+            ]
+        }
+    },
+    {
+        "id": 26340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003506.wav"
+            ]
+        }
+    },
+    {
+        "id": 26342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001303.wav"
+            ]
+        }
+    },
+    {
+        "id": 26348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001635.wav"
+            ]
+        }
+    },
+    {
+        "id": 26353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003679.wav"
+            ]
+        }
+    },
+    {
+        "id": 26356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004018.wav"
+            ]
+        }
+    },
+    {
+        "id": 26357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001739.wav"
+            ]
+        }
+    },
+    {
+        "id": 26363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004190.wav"
+            ]
+        }
+    },
+    {
+        "id": 26365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001481.wav"
+            ]
+        }
+    },
+    {
+        "id": 26369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008114.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008114.wav"
+            ]
+        }
+    },
+    {
+        "id": 26378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004189.wav"
+            ]
+        }
+    },
+    {
+        "id": 26380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004392.wav"
+            ]
+        }
+    },
+    {
+        "id": 26395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005492.wav"
+            ]
+        }
+    },
+    {
+        "id": 26400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007329.wav"
+            ]
+        }
+    },
+    {
+        "id": 26412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005482.wav"
+            ]
+        }
+    },
+    {
+        "id": 26416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002419.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002419.wav"
+            ]
+        }
+    },
+    {
+        "id": 26417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the pipa on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006043.wav"
+            ]
+        }
+    },
+    {
+        "id": 26423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006730.wav"
+            ]
+        }
+    },
+    {
+        "id": 26426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005406.wav"
+            ]
+        }
+    },
+    {
+        "id": 26430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000571.wav"
+            ]
+        }
+    },
+    {
+        "id": 26432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004824.wav"
+            ]
+        }
+    },
+    {
+        "id": 26435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003764.wav"
+            ]
+        }
+    },
+    {
+        "id": 26436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002463.wav"
+            ]
+        }
+    },
+    {
+        "id": 26437,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008903.wav"
+            ]
+        }
+    },
+    {
+        "id": 26448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002004.wav"
+            ]
+        }
+    },
+    {
+        "id": 26456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000839.wav"
+            ]
+        }
+    },
+    {
+        "id": 26465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004169.wav"
+            ]
+        }
+    },
+    {
+        "id": 26467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008581.wav"
+            ]
+        }
+    },
+    {
+        "id": 26470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004017.wav"
+            ]
+        }
+    },
+    {
+        "id": 26478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004306.wav"
+            ]
+        }
+    },
+    {
+        "id": 26480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001375.wav"
+            ]
+        }
+    },
+    {
+        "id": 26481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005953.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005953.wav"
+            ]
+        }
+    },
+    {
+        "id": 26482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet playing longer than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005512.wav"
+            ]
+        }
+    },
+    {
+        "id": 26484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008530.wav"
+            ]
+        }
+    },
+    {
+        "id": 26487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000240.wav"
+            ]
+        }
+    },
+    {
+        "id": 26489,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and erhu sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002519.wav"
+            ]
+        }
+    },
+    {
+        "id": 26497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004892.wav"
+            ]
+        }
+    },
+    {
+        "id": 26499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006784.wav"
+            ]
+        }
+    },
+    {
+        "id": 26506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008464.wav"
+            ]
+        }
+    },
+    {
+        "id": 26511,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005927.wav"
+            ]
+        }
+    },
+    {
+        "id": 26522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001541.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001541.wav"
+            ]
+        }
+    },
+    {
+        "id": 26527,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008290.wav"
+            ]
+        }
+    },
+    {
+        "id": 26533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006193.wav"
+            ]
+        }
+    },
+    {
+        "id": 26535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004848.wav"
+            ]
+        }
+    },
+    {
+        "id": 26549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004638.wav"
+            ]
+        }
+    },
+    {
+        "id": 26550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004759.wav"
+            ]
+        }
+    },
+    {
+        "id": 26551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005016.wav"
+            ]
+        }
+    },
+    {
+        "id": 26556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000769.wav"
+            ]
+        }
+    },
+    {
+        "id": 26567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007451.wav"
+            ]
+        }
+    },
+    {
+        "id": 26581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008784.wav"
+            ]
+        }
+    },
+    {
+        "id": 26590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004834.wav"
+            ]
+        }
+    },
+    {
+        "id": 26603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004519.wav"
+            ]
+        }
+    },
+    {
+        "id": 26607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004721.wav"
+            ]
+        }
+    },
+    {
+        "id": 26611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002110.wav"
+            ]
+        }
+    },
+    {
+        "id": 26613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002071.wav"
+            ]
+        }
+    },
+    {
+        "id": 26615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008200.wav"
+            ]
+        }
+    },
+    {
+        "id": 26617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002067.wav"
+            ]
+        }
+    },
+    {
+        "id": 26623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005732.wav"
+            ]
+        }
+    },
+    {
+        "id": 26624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005469.wav"
+            ]
+        }
+    },
+    {
+        "id": 26638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004252.wav"
+            ]
+        }
+    },
+    {
+        "id": 26646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008137.wav"
+            ]
+        }
+    },
+    {
+        "id": 26657,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002116.wav"
+            ]
+        }
+    },
+    {
+        "id": 26659,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004139.wav"
+            ]
+        }
+    },
+    {
+        "id": 26661,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004450.wav"
+            ]
+        }
+    },
+    {
+        "id": 26665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000427.wav"
+            ]
+        }
+    },
+    {
+        "id": 26667,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008430.wav"
+            ]
+        }
+    },
+    {
+        "id": 26683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004796.wav"
+            ]
+        }
+    },
+    {
+        "id": 26686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005779.wav"
+            ]
+        }
+    },
+    {
+        "id": 26721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001262.wav"
+            ]
+        }
+    },
+    {
+        "id": 26725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002524.wav"
+            ]
+        }
+    },
+    {
+        "id": 26730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003922.wav"
+            ]
+        }
+    },
+    {
+        "id": 26734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002939.wav"
+            ]
+        }
+    },
+    {
+        "id": 26737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004381.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004381.wav"
+            ]
+        }
+    },
+    {
+        "id": 26765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005744.wav"
+            ]
+        }
+    },
+    {
+        "id": 26767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001818.wav"
+            ]
+        }
+    },
+    {
+        "id": 26768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000592.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000592.wav"
+            ]
+        }
+    },
+    {
+        "id": 26771,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001962.wav"
+            ]
+        }
+    },
+    {
+        "id": 26786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006786.wav"
+            ]
+        }
+    },
+    {
+        "id": 26787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007437.wav"
+            ]
+        }
+    },
+    {
+        "id": 26800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007249.wav"
+            ]
+        }
+    },
+    {
+        "id": 26805,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004450.wav"
+            ]
+        }
+    },
+    {
+        "id": 26826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006031.wav"
+            ]
+        }
+    },
+    {
+        "id": 26832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008681.wav"
+            ]
+        }
+    },
+    {
+        "id": 26838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005760.wav"
+            ]
+        }
+    },
+    {
+        "id": 26839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003895.wav"
+            ]
+        }
+    },
+    {
+        "id": 26848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007620.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007620.wav"
+            ]
+        }
+    },
+    {
+        "id": 26849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001733.wav"
+            ]
+        }
+    },
+    {
+        "id": 26850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002192.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002192.wav"
+            ]
+        }
+    },
+    {
+        "id": 26855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003963.wav"
+            ]
+        }
+    },
+    {
+        "id": 26858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006222.wav"
+            ]
+        }
+    },
+    {
+        "id": 26860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001727.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001727.wav"
+            ]
+        }
+    },
+    {
+        "id": 26861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000317.wav"
+            ]
+        }
+    },
+    {
+        "id": 26862,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001563.wav"
+            ]
+        }
+    },
+    {
+        "id": 26864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004552.wav"
+            ]
+        }
+    },
+    {
+        "id": 26867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008561.wav"
+            ]
+        }
+    },
+    {
+        "id": 26908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002346.wav"
+            ]
+        }
+    },
+    {
+        "id": 26909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas on the left louder than the congas on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000253.wav"
+            ]
+        }
+    },
+    {
+        "id": 26911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006934.wav"
+            ]
+        }
+    },
+    {
+        "id": 26918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006624.wav"
+            ]
+        }
+    },
+    {
+        "id": 26921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008438.wav"
+            ]
+        }
+    },
+    {
+        "id": 26922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005159.wav"
+            ]
+        }
+    },
+    {
+        "id": 26925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the middle louder than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007387.wav"
+            ]
+        }
+    },
+    {
+        "id": 26933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001420.wav"
+            ]
+        }
+    },
+    {
+        "id": 26936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002248.wav"
+            ]
+        }
+    },
+    {
+        "id": 26939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007460.wav"
+            ]
+        }
+    },
+    {
+        "id": 26945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002512.wav"
+            ]
+        }
+    },
+    {
+        "id": 26962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001400.wav"
+            ]
+        }
+    },
+    {
+        "id": 26970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006428.wav"
+            ]
+        }
+    },
+    {
+        "id": 26973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003307.wav"
+            ]
+        }
+    },
+    {
+        "id": 26975,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003632.wav"
+            ]
+        }
+    },
+    {
+        "id": 26990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002123.wav"
+            ]
+        }
+    },
+    {
+        "id": 26991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003321.wav"
+            ]
+        }
+    },
+    {
+        "id": 26996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007528.wav"
+            ]
+        }
+    },
+    {
+        "id": 26997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002996.wav"
+            ]
+        }
+    },
+    {
+        "id": 26998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005237.wav"
+            ]
+        }
+    },
+    {
+        "id": 27001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005214.wav"
+            ]
+        }
+    },
+    {
+        "id": 27022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006922.wav"
+            ]
+        }
+    },
+    {
+        "id": 27023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007379.wav"
+            ]
+        }
+    },
+    {
+        "id": 27025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000540.wav"
+            ]
+        }
+    },
+    {
+        "id": 27028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007744.wav"
+            ]
+        }
+    },
+    {
+        "id": 27039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005585.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005585.wav"
+            ]
+        }
+    },
+    {
+        "id": 27042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006288.wav"
+            ]
+        }
+    },
+    {
+        "id": 27044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002574.wav"
+            ]
+        }
+    },
+    {
+        "id": 27046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000661.wav"
+            ]
+        }
+    },
+    {
+        "id": 27054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003128.wav"
+            ]
+        }
+    },
+    {
+        "id": 27063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006351.wav"
+            ]
+        }
+    },
+    {
+        "id": 27064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004180.wav"
+            ]
+        }
+    },
+    {
+        "id": 27069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000504.wav"
+            ]
+        }
+    },
+    {
+        "id": 27070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003735.wav"
+            ]
+        }
+    },
+    {
+        "id": 27084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000285.wav"
+            ]
+        }
+    },
+    {
+        "id": 27093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007891.wav"
+            ]
+        }
+    },
+    {
+        "id": 27107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004688.wav"
+            ]
+        }
+    },
+    {
+        "id": 27110,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007744.wav"
+            ]
+        }
+    },
+    {
+        "id": 27111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008682.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008682.wav"
+            ]
+        }
+    },
+    {
+        "id": 27116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005997.wav"
+            ]
+        }
+    },
+    {
+        "id": 27117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008754.wav"
+            ]
+        }
+    },
+    {
+        "id": 27118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001574.wav"
+            ]
+        }
+    },
+    {
+        "id": 27119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001471.wav"
+            ]
+        }
+    },
+    {
+        "id": 27129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005609.wav"
+            ]
+        }
+    },
+    {
+        "id": 27132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000514.wav"
+            ]
+        }
+    },
+    {
+        "id": 27133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and xylophone sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005685.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005685.wav"
+            ]
+        }
+    },
+    {
+        "id": 27138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and banjo sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006390.wav"
+            ]
+        }
+    },
+    {
+        "id": 27139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004023.wav"
+            ]
+        }
+    },
+    {
+        "id": 27141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002172.wav"
+            ]
+        }
+    },
+    {
+        "id": 27156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004707.wav"
+            ]
+        }
+    },
+    {
+        "id": 27162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006639.wav"
+            ]
+        }
+    },
+    {
+        "id": 27172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001414.wav"
+            ]
+        }
+    },
+    {
+        "id": 27175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008747.wav"
+            ]
+        }
+    },
+    {
+        "id": 27177,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003874.wav"
+            ]
+        }
+    },
+    {
+        "id": 27180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002101.wav"
+            ]
+        }
+    },
+    {
+        "id": 27184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001555.wav"
+            ]
+        }
+    },
+    {
+        "id": 27187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008887.wav"
+            ]
+        }
+    },
+    {
+        "id": 27189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003639.wav"
+            ]
+        }
+    },
+    {
+        "id": 27205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004852.wav"
+            ]
+        }
+    },
+    {
+        "id": 27214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003250.wav"
+            ]
+        }
+    },
+    {
+        "id": 27216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003869.wav"
+            ]
+        }
+    },
+    {
+        "id": 27224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008339.wav"
+            ]
+        }
+    },
+    {
+        "id": 27228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008169.wav"
+            ]
+        }
+    },
+    {
+        "id": 27231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007487.wav"
+            ]
+        }
+    },
+    {
+        "id": 27247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005190.wav"
+            ]
+        }
+    },
+    {
+        "id": 27248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001342.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001342.wav"
+            ]
+        }
+    },
+    {
+        "id": 27249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006064.wav"
+            ]
+        }
+    },
+    {
+        "id": 27250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there pipa and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006510.wav"
+            ]
+        }
+    },
+    {
+        "id": 27251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002305.wav"
+            ]
+        }
+    },
+    {
+        "id": 27252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003706.wav"
+            ]
+        }
+    },
+    {
+        "id": 27254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002047.wav"
+            ]
+        }
+    },
+    {
+        "id": 27258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002090.wav"
+            ]
+        }
+    },
+    {
+        "id": 27259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002444.wav"
+            ]
+        }
+    },
+    {
+        "id": 27264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000888.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000888.wav"
+            ]
+        }
+    },
+    {
+        "id": 27270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006661.wav"
+            ]
+        }
+    },
+    {
+        "id": 27272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and cello sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007649.wav"
+            ]
+        }
+    },
+    {
+        "id": 27279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002380.wav"
+            ]
+        }
+    },
+    {
+        "id": 27280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000640.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000640.wav"
+            ]
+        }
+    },
+    {
+        "id": 27282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006441.wav"
+            ]
+        }
+    },
+    {
+        "id": 27284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and trumpet sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007086.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007086.wav"
+            ]
+        }
+    },
+    {
+        "id": 27287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004994.wav"
+            ]
+        }
+    },
+    {
+        "id": 27289,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and flute instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002365.wav"
+            ]
+        }
+    },
+    {
+        "id": 27300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003763.wav"
+            ]
+        }
+    },
+    {
+        "id": 27310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004380.wav"
+            ]
+        }
+    },
+    {
+        "id": 27314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006950.wav"
+            ]
+        }
+    },
+    {
+        "id": 27320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007568.wav"
+            ]
+        }
+    },
+    {
+        "id": 27324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006467.wav"
+            ]
+        }
+    },
+    {
+        "id": 27325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005269.wav"
+            ]
+        }
+    },
+    {
+        "id": 27345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006735.wav"
+            ]
+        }
+    },
+    {
+        "id": 27357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004084.wav"
+            ]
+        }
+    },
+    {
+        "id": 27367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004224.wav"
+            ]
+        }
+    },
+    {
+        "id": 27377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002625.wav"
+            ]
+        }
+    },
+    {
+        "id": 27380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007353.wav"
+            ]
+        }
+    },
+    {
+        "id": 27381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004078.wav"
+            ]
+        }
+    },
+    {
+        "id": 27382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003968.wav"
+            ]
+        }
+    },
+    {
+        "id": 27386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006866.wav"
+            ]
+        }
+    },
+    {
+        "id": 27390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004944.wav"
+            ]
+        }
+    },
+    {
+        "id": 27393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006761.wav"
+            ]
+        }
+    },
+    {
+        "id": 27402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002999.wav"
+            ]
+        }
+    },
+    {
+        "id": 27406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001157.wav"
+            ]
+        }
+    },
+    {
+        "id": 27419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006999.wav"
+            ]
+        }
+    },
+    {
+        "id": 27422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006739.wav"
+            ]
+        }
+    },
+    {
+        "id": 27441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000546.wav"
+            ]
+        }
+    },
+    {
+        "id": 27443,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003660.wav"
+            ]
+        }
+    },
+    {
+        "id": 27445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008628.wav"
+            ]
+        }
+    },
+    {
+        "id": 27456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007660.wav"
+            ]
+        }
+    },
+    {
+        "id": 27460,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001517.wav"
+            ]
+        }
+    },
+    {
+        "id": 27462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007293.wav"
+            ]
+        }
+    },
+    {
+        "id": 27463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007241.wav"
+            ]
+        }
+    },
+    {
+        "id": 27475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004091.wav"
+            ]
+        }
+    },
+    {
+        "id": 27476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002677.wav"
+            ]
+        }
+    },
+    {
+        "id": 27478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004632.wav"
+            ]
+        }
+    },
+    {
+        "id": 27505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001036.wav"
+            ]
+        }
+    },
+    {
+        "id": 27509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008155.wav"
+            ]
+        }
+    },
+    {
+        "id": 27518,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002617.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002617.wav"
+            ]
+        }
+    },
+    {
+        "id": 27523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002683.wav"
+            ]
+        }
+    },
+    {
+        "id": 27529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005744.wav"
+            ]
+        }
+    },
+    {
+        "id": 27533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003119.wav"
+            ]
+        }
+    },
+    {
+        "id": 27538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and tuba sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000967.wav"
+            ]
+        }
+    },
+    {
+        "id": 27543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004301.wav"
+            ]
+        }
+    },
+    {
+        "id": 27548,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005958.wav"
+            ]
+        }
+    },
+    {
+        "id": 27550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001901.wav"
+            ]
+        }
+    },
+    {
+        "id": 27557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003326.wav"
+            ]
+        }
+    },
+    {
+        "id": 27558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004536.wav"
+            ]
+        }
+    },
+    {
+        "id": 27561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007105.wav"
+            ]
+        }
+    },
+    {
+        "id": 27563,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs there a voiceover?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002029.wav"
+            ]
+        }
+    },
+    {
+        "id": 27568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002506.wav"
+            ]
+        }
+    },
+    {
+        "id": 27570,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004733.wav"
+            ]
+        }
+    },
+    {
+        "id": 27572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004375.wav"
+            ]
+        }
+    },
+    {
+        "id": 27574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005815.wav"
+            ]
+        }
+    },
+    {
+        "id": 27578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and violin sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004891.wav"
+            ]
+        }
+    },
+    {
+        "id": 27589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000755.wav"
+            ]
+        }
+    },
+    {
+        "id": 27597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002839.wav"
+            ]
+        }
+    },
+    {
+        "id": 27599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and piano sound?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005766.wav"
+            ]
+        }
+    },
+    {
+        "id": 27601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008706.wav"
+            ]
+        }
+    },
+    {
+        "id": 27604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004875.wav"
+            ]
+        }
+    },
+    {
+        "id": 27606,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007098.wav"
+            ]
+        }
+    },
+    {
+        "id": 27609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007349.wav"
+            ]
+        }
+    },
+    {
+        "id": 27613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007239.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007239.wav"
+            ]
+        }
+    },
+    {
+        "id": 27615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005613.wav"
+            ]
+        }
+    },
+    {
+        "id": 27625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002194.wav"
+            ]
+        }
+    },
+    {
+        "id": 27628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007689.wav"
+            ]
+        }
+    },
+    {
+        "id": 27630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004549.wav"
+            ]
+        }
+    },
+    {
+        "id": 27634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001905.wav"
+            ]
+        }
+    },
+    {
+        "id": 27635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001527.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001527.wav"
+            ]
+        }
+    },
+    {
+        "id": 27644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004348.wav"
+            ]
+        }
+    },
+    {
+        "id": 27655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006700.wav"
+            ]
+        }
+    },
+    {
+        "id": 27671,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006730.wav"
+            ]
+        }
+    },
+    {
+        "id": 27676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004568.wav"
+            ]
+        }
+    },
+    {
+        "id": 27678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002550.wav"
+            ]
+        }
+    },
+    {
+        "id": 27685,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007284.wav"
+            ]
+        }
+    },
+    {
+        "id": 27694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008722.wav"
+            ]
+        }
+    },
+    {
+        "id": 27702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005323.wav"
+            ]
+        }
+    },
+    {
+        "id": 27712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004571.wav"
+            ]
+        }
+    },
+    {
+        "id": 27726,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001068.wav"
+            ]
+        }
+    },
+    {
+        "id": 27728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003684.wav"
+            ]
+        }
+    },
+    {
+        "id": 27730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005104.wav"
+            ]
+        }
+    },
+    {
+        "id": 27742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001958.wav"
+            ]
+        }
+    },
+    {
+        "id": 27753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008613.wav"
+            ]
+        }
+    },
+    {
+        "id": 27757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004066.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004066.wav"
+            ]
+        }
+    },
+    {
+        "id": 27758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006009.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006009.wav"
+            ]
+        }
+    },
+    {
+        "id": 27762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008026.wav"
+            ]
+        }
+    },
+    {
+        "id": 27768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003327.wav"
+            ]
+        }
+    },
+    {
+        "id": 27775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005106.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005106.wav"
+            ]
+        }
+    },
+    {
+        "id": 27779,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000853.wav"
+            ]
+        }
+    },
+    {
+        "id": 27780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002990.wav"
+            ]
+        }
+    },
+    {
+        "id": 27781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000601.wav"
+            ]
+        }
+    },
+    {
+        "id": 27790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and pipa instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002825.wav"
+            ]
+        }
+    },
+    {
+        "id": 27799,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002198.wav"
+            ]
+        }
+    },
+    {
+        "id": 27809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich acoustic_guitar makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002397.wav"
+            ]
+        }
+    },
+    {
+        "id": 27813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003918.wav"
+            ]
+        }
+    },
+    {
+        "id": 27814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001188.wav"
+            ]
+        }
+    },
+    {
+        "id": 27822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000602.wav"
+            ]
+        }
+    },
+    {
+        "id": 27832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000659.wav"
+            ]
+        }
+    },
+    {
+        "id": 27839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000078.wav"
+            ]
+        }
+    },
+    {
+        "id": 27840,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004578.wav"
+            ]
+        }
+    },
+    {
+        "id": 27846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008949.wav"
+            ]
+        }
+    },
+    {
+        "id": 27852,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006181.wav"
+            ]
+        }
+    },
+    {
+        "id": 27859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007733.wav"
+            ]
+        }
+    },
+    {
+        "id": 27875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000450.wav"
+            ]
+        }
+    },
+    {
+        "id": 27876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001731.wav"
+            ]
+        }
+    },
+    {
+        "id": 27896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000761.wav"
+            ]
+        }
+    },
+    {
+        "id": 27898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001180.wav"
+            ]
+        }
+    },
+    {
+        "id": 27909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003744.wav"
+            ]
+        }
+    },
+    {
+        "id": 27912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007844.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007844.wav"
+            ]
+        }
+    },
+    {
+        "id": 27915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and xylophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007088.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007088.wav"
+            ]
+        }
+    },
+    {
+        "id": 27921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and bassoon instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005834.wav"
+            ]
+        }
+    },
+    {
+        "id": 27927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001228.wav"
+            ]
+        }
+    },
+    {
+        "id": 27932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004445.wav"
+            ]
+        }
+    },
+    {
+        "id": 27934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005053.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005053.wav"
+            ]
+        }
+    },
+    {
+        "id": 27936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006630.wav"
+            ]
+        }
+    },
+    {
+        "id": 27939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004221.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004221.wav"
+            ]
+        }
+    },
+    {
+        "id": 27940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008962.wav"
+            ]
+        }
+    },
+    {
+        "id": 27945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008090.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008090.wav"
+            ]
+        }
+    },
+    {
+        "id": 27948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004083.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004083.wav"
+            ]
+        }
+    },
+    {
+        "id": 27954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003840.wav"
+            ]
+        }
+    },
+    {
+        "id": 27976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007082.wav"
+            ]
+        }
+    },
+    {
+        "id": 27981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005123.wav"
+            ]
+        }
+    },
+    {
+        "id": 27988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007110.wav"
+            ]
+        }
+    },
+    {
+        "id": 27989,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002480.wav"
+            ]
+        }
+    },
+    {
+        "id": 27990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005989.wav"
+            ]
+        }
+    },
+    {
+        "id": 27992,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007773.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007773.wav"
+            ]
+        }
+    },
+    {
+        "id": 28006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001850.wav"
+            ]
+        }
+    },
+    {
+        "id": 28019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007894.wav"
+            ]
+        }
+    },
+    {
+        "id": 28031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and congas instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004765.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004765.wav"
+            ]
+        }
+    },
+    {
+        "id": 28034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001244.wav"
+            ]
+        }
+    },
+    {
+        "id": 28039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003960.wav"
+            ]
+        }
+    },
+    {
+        "id": 28047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005409.wav"
+            ]
+        }
+    },
+    {
+        "id": 28048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006251.wav"
+            ]
+        }
+    },
+    {
+        "id": 28054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006423.wav"
+            ]
+        }
+    },
+    {
+        "id": 28062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002606.wav"
+            ]
+        }
+    },
+    {
+        "id": 28063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and electric_bass instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005319.wav"
+            ]
+        }
+    },
+    {
+        "id": 28068,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and congas instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003930.wav"
+            ]
+        }
+    },
+    {
+        "id": 28073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and congas instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001615.wav"
+            ]
+        }
+    },
+    {
+        "id": 28076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008283.wav"
+            ]
+        }
+    },
+    {
+        "id": 28088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007067.wav"
+            ]
+        }
+    },
+    {
+        "id": 28091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002126.wav"
+            ]
+        }
+    },
+    {
+        "id": 28097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007153.wav"
+            ]
+        }
+    },
+    {
+        "id": 28099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006001.wav"
+            ]
+        }
+    },
+    {
+        "id": 28103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004749.wav"
+            ]
+        }
+    },
+    {
+        "id": 28131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007781.wav"
+            ]
+        }
+    },
+    {
+        "id": 28133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003162.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003162.wav"
+            ]
+        }
+    },
+    {
+        "id": 28144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002688.wav"
+            ]
+        }
+    },
+    {
+        "id": 28145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002754.wav"
+            ]
+        }
+    },
+    {
+        "id": 28146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007573.wav"
+            ]
+        }
+    },
+    {
+        "id": 28152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005701.wav"
+            ]
+        }
+    },
+    {
+        "id": 28154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003871.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003871.wav"
+            ]
+        }
+    },
+    {
+        "id": 28155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003905.wav"
+            ]
+        }
+    },
+    {
+        "id": 28156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001825.wav"
+            ]
+        }
+    },
+    {
+        "id": 28166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004096.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004096.wav"
+            ]
+        }
+    },
+    {
+        "id": 28168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001988.wav"
+            ]
+        }
+    },
+    {
+        "id": 28169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000123.wav"
+            ]
+        }
+    },
+    {
+        "id": 28170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005531.wav"
+            ]
+        }
+    },
+    {
+        "id": 28174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008316.wav"
+            ]
+        }
+    },
+    {
+        "id": 28176,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003568.wav"
+            ]
+        }
+    },
+    {
+        "id": 28179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006016.wav"
+            ]
+        }
+    },
+    {
+        "id": 28180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004581.wav"
+            ]
+        }
+    },
+    {
+        "id": 28182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006347.wav"
+            ]
+        }
+    },
+    {
+        "id": 28184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006892.wav"
+            ]
+        }
+    },
+    {
+        "id": 28190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005481.wav"
+            ]
+        }
+    },
+    {
+        "id": 28191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002428.wav"
+            ]
+        }
+    },
+    {
+        "id": 28199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008422.wav"
+            ]
+        }
+    },
+    {
+        "id": 28203,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000073.wav"
+            ]
+        }
+    },
+    {
+        "id": 28204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002995.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002995.wav"
+            ]
+        }
+    },
+    {
+        "id": 28227,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000147.wav"
+            ]
+        }
+    },
+    {
+        "id": 28237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000249.wav"
+            ]
+        }
+    },
+    {
+        "id": 28239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007218.wav"
+            ]
+        }
+    },
+    {
+        "id": 28249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002019.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002019.wav"
+            ]
+        }
+    },
+    {
+        "id": 28250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003838.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003838.wav"
+            ]
+        }
+    },
+    {
+        "id": 28255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007145.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007145.wav"
+            ]
+        }
+    },
+    {
+        "id": 28259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008764.wav"
+            ]
+        }
+    },
+    {
+        "id": 28261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001012.wav"
+            ]
+        }
+    },
+    {
+        "id": 28265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001047.wav"
+            ]
+        }
+    },
+    {
+        "id": 28266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004904.wav"
+            ]
+        }
+    },
+    {
+        "id": 28271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006741.wav"
+            ]
+        }
+    },
+    {
+        "id": 28272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003582.wav"
+            ]
+        }
+    },
+    {
+        "id": 28281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004958.wav"
+            ]
+        }
+    },
+    {
+        "id": 28283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005021.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005021.wav"
+            ]
+        }
+    },
+    {
+        "id": 28290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008935.wav"
+            ]
+        }
+    },
+    {
+        "id": 28301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007675.wav"
+            ]
+        }
+    },
+    {
+        "id": 28305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002095.wav"
+            ]
+        }
+    },
+    {
+        "id": 28310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003439.wav"
+            ]
+        }
+    },
+    {
+        "id": 28315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001579.wav"
+            ]
+        }
+    },
+    {
+        "id": 28334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008636.wav"
+            ]
+        }
+    },
+    {
+        "id": 28335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005336.wav"
+            ]
+        }
+    },
+    {
+        "id": 28336,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002883.wav"
+            ]
+        }
+    },
+    {
+        "id": 28337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002268.wav"
+            ]
+        }
+    },
+    {
+        "id": 28340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000177.wav"
+            ]
+        }
+    },
+    {
+        "id": 28343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006746.wav"
+            ]
+        }
+    },
+    {
+        "id": 28345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008863.wav"
+            ]
+        }
+    },
+    {
+        "id": 28347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008442.wav"
+            ]
+        }
+    },
+    {
+        "id": 28356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006949.wav"
+            ]
+        }
+    },
+    {
+        "id": 28363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003704.wav"
+            ]
+        }
+    },
+    {
+        "id": 28367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004961.wav"
+            ]
+        }
+    },
+    {
+        "id": 28368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001816.wav"
+            ]
+        }
+    },
+    {
+        "id": 28370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003099.wav"
+            ]
+        }
+    },
+    {
+        "id": 28375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000779.wav"
+            ]
+        }
+    },
+    {
+        "id": 28382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004425.wav"
+            ]
+        }
+    },
+    {
+        "id": 28385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002575.wav"
+            ]
+        }
+    },
+    {
+        "id": 28386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008002.wav"
+            ]
+        }
+    },
+    {
+        "id": 28391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000049.wav"
+            ]
+        }
+    },
+    {
+        "id": 28396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006534.wav"
+            ]
+        }
+    },
+    {
+        "id": 28404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004916.wav"
+            ]
+        }
+    },
+    {
+        "id": 28405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005381.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005381.wav"
+            ]
+        }
+    },
+    {
+        "id": 28414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001918.wav"
+            ]
+        }
+    },
+    {
+        "id": 28417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006075.wav"
+            ]
+        }
+    },
+    {
+        "id": 28422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007089.wav"
+            ]
+        }
+    },
+    {
+        "id": 28425,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005874.wav"
+            ]
+        }
+    },
+    {
+        "id": 28436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003524.wav"
+            ]
+        }
+    },
+    {
+        "id": 28439,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004071.wav"
+            ]
+        }
+    },
+    {
+        "id": 28448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007601.wav"
+            ]
+        }
+    },
+    {
+        "id": 28453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000017.wav"
+            ]
+        }
+    },
+    {
+        "id": 28456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003300.wav"
+            ]
+        }
+    },
+    {
+        "id": 28461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006660.wav"
+            ]
+        }
+    },
+    {
+        "id": 28465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006615.wav"
+            ]
+        }
+    },
+    {
+        "id": 28467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008263.wav"
+            ]
+        }
+    },
+    {
+        "id": 28476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001285.wav"
+            ]
+        }
+    },
+    {
+        "id": 28477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008686.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008686.wav"
+            ]
+        }
+    },
+    {
+        "id": 28479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003816.wav"
+            ]
+        }
+    },
+    {
+        "id": 28486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000496.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000496.wav"
+            ]
+        }
+    },
+    {
+        "id": 28487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001839.wav"
+            ]
+        }
+    },
+    {
+        "id": 28504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007918.wav"
+            ]
+        }
+    },
+    {
+        "id": 28506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006220.wav"
+            ]
+        }
+    },
+    {
+        "id": 28509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000984.wav"
+            ]
+        }
+    },
+    {
+        "id": 28513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003017.wav"
+            ]
+        }
+    },
+    {
+        "id": 28517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008590.wav"
+            ]
+        }
+    },
+    {
+        "id": 28519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001085.wav"
+            ]
+        }
+    },
+    {
+        "id": 28520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005026.wav"
+            ]
+        }
+    },
+    {
+        "id": 28521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000387.wav"
+            ]
+        }
+    },
+    {
+        "id": 28533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002648.wav"
+            ]
+        }
+    },
+    {
+        "id": 28534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006577.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006577.wav"
+            ]
+        }
+    },
+    {
+        "id": 28551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001705.wav"
+            ]
+        }
+    },
+    {
+        "id": 28561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003668.wav"
+            ]
+        }
+    },
+    {
+        "id": 28562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001380.wav"
+            ]
+        }
+    },
+    {
+        "id": 28576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005099.wav"
+            ]
+        }
+    },
+    {
+        "id": 28578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007798.wav"
+            ]
+        }
+    },
+    {
+        "id": 28580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005545.wav"
+            ]
+        }
+    },
+    {
+        "id": 28584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008618.wav"
+            ]
+        }
+    },
+    {
+        "id": 28589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002614.wav"
+            ]
+        }
+    },
+    {
+        "id": 28597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006845.wav"
+            ]
+        }
+    },
+    {
+        "id": 28599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006761.wav"
+            ]
+        }
+    },
+    {
+        "id": 28604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003851.wav"
+            ]
+        }
+    },
+    {
+        "id": 28605,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006849.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006849.wav"
+            ]
+        }
+    },
+    {
+        "id": 28608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004578.wav"
+            ]
+        }
+    },
+    {
+        "id": 28611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006980.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006980.wav"
+            ]
+        }
+    },
+    {
+        "id": 28612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001882.wav"
+            ]
+        }
+    },
+    {
+        "id": 28614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006043.wav"
+            ]
+        }
+    },
+    {
+        "id": 28616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003061.wav"
+            ]
+        }
+    },
+    {
+        "id": 28618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003370.wav"
+            ]
+        }
+    },
+    {
+        "id": 28619,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002643.wav"
+            ]
+        }
+    },
+    {
+        "id": 28621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000975.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000975.wav"
+            ]
+        }
+    },
+    {
+        "id": 28623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005863.wav"
+            ]
+        }
+    },
+    {
+        "id": 28629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004216.wav"
+            ]
+        }
+    },
+    {
+        "id": 28631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008288.wav"
+            ]
+        }
+    },
+    {
+        "id": 28632,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003932.wav"
+            ]
+        }
+    },
+    {
+        "id": 28637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004474.wav"
+            ]
+        }
+    },
+    {
+        "id": 28642,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001111.wav"
+            ]
+        }
+    },
+    {
+        "id": 28649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002054.wav"
+            ]
+        }
+    },
+    {
+        "id": 28653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001971.wav"
+            ]
+        }
+    },
+    {
+        "id": 28655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008963.wav"
+            ]
+        }
+    },
+    {
+        "id": 28656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006038.wav"
+            ]
+        }
+    },
+    {
+        "id": 28663,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006958.wav"
+            ]
+        }
+    },
+    {
+        "id": 28665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005630.wav"
+            ]
+        }
+    },
+    {
+        "id": 28668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004288.wav"
+            ]
+        }
+    },
+    {
+        "id": 28672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and congas instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001034.wav"
+            ]
+        }
+    },
+    {
+        "id": 28673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003253.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003253.wav"
+            ]
+        }
+    },
+    {
+        "id": 28677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007179.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007179.wav"
+            ]
+        }
+    },
+    {
+        "id": 28678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003155.wav"
+            ]
+        }
+    },
+    {
+        "id": 28680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001228.wav"
+            ]
+        }
+    },
+    {
+        "id": 28681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and flute instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002521.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002521.wav"
+            ]
+        }
+    },
+    {
+        "id": 28690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008840.wav"
+            ]
+        }
+    },
+    {
+        "id": 28694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005525.wav"
+            ]
+        }
+    },
+    {
+        "id": 28698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008183.wav"
+            ]
+        }
+    },
+    {
+        "id": 28699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008115.wav"
+            ]
+        }
+    },
+    {
+        "id": 28703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006910.wav"
+            ]
+        }
+    },
+    {
+        "id": 28705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005053.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005053.wav"
+            ]
+        }
+    },
+    {
+        "id": 28709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and suona instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007144.wav"
+            ]
+        }
+    },
+    {
+        "id": 28717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006029.wav"
+            ]
+        }
+    },
+    {
+        "id": 28723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001840.wav"
+            ]
+        }
+    },
+    {
+        "id": 28729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007408.wav"
+            ]
+        }
+    },
+    {
+        "id": 28739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007199.wav"
+            ]
+        }
+    },
+    {
+        "id": 28742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001145.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001145.wav"
+            ]
+        }
+    },
+    {
+        "id": 28745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007334.wav"
+            ]
+        }
+    },
+    {
+        "id": 28747,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000174.wav"
+            ]
+        }
+    },
+    {
+        "id": 28758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002138.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002138.wav"
+            ]
+        }
+    },
+    {
+        "id": 28771,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005087.wav"
+            ]
+        }
+    },
+    {
+        "id": 28781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001532.wav"
+            ]
+        }
+    },
+    {
+        "id": 28783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005396.wav"
+            ]
+        }
+    },
+    {
+        "id": 28789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003754.wav"
+            ]
+        }
+    },
+    {
+        "id": 28802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004976.wav"
+            ]
+        }
+    },
+    {
+        "id": 28808,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003479.wav"
+            ]
+        }
+    },
+    {
+        "id": 28811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001762.wav"
+            ]
+        }
+    },
+    {
+        "id": 28817,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004104.wav"
+            ]
+        }
+    },
+    {
+        "id": 28820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008102.wav"
+            ]
+        }
+    },
+    {
+        "id": 28824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008125.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008125.wav"
+            ]
+        }
+    },
+    {
+        "id": 28838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001508.wav"
+            ]
+        }
+    },
+    {
+        "id": 28842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001136.wav"
+            ]
+        }
+    },
+    {
+        "id": 28844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005631.wav"
+            ]
+        }
+    },
+    {
+        "id": 28848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003685.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003685.wav"
+            ]
+        }
+    },
+    {
+        "id": 28851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003039.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003039.wav"
+            ]
+        }
+    },
+    {
+        "id": 28855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003821.wav"
+            ]
+        }
+    },
+    {
+        "id": 28857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002238.wav"
+            ]
+        }
+    },
+    {
+        "id": 28864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001549.wav"
+            ]
+        }
+    },
+    {
+        "id": 28872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004321.wav"
+            ]
+        }
+    },
+    {
+        "id": 28873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there drum and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006723.wav"
+            ]
+        }
+    },
+    {
+        "id": 28876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007439.wav"
+            ]
+        }
+    },
+    {
+        "id": 28883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005642.wav"
+            ]
+        }
+    },
+    {
+        "id": 28884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007434.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007434.wav"
+            ]
+        }
+    },
+    {
+        "id": 28894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001463.wav"
+            ]
+        }
+    },
+    {
+        "id": 28895,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003310.wav"
+            ]
+        }
+    },
+    {
+        "id": 28897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002191.wav"
+            ]
+        }
+    },
+    {
+        "id": 28904,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and drum instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002211.wav"
+            ]
+        }
+    },
+    {
+        "id": 28909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002316.wav"
+            ]
+        }
+    },
+    {
+        "id": 28923,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006492.wav"
+            ]
+        }
+    },
+    {
+        "id": 28925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000315.wav"
+            ]
+        }
+    },
+    {
+        "id": 28933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008214.wav"
+            ]
+        }
+    },
+    {
+        "id": 28936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005337.wav"
+            ]
+        }
+    },
+    {
+        "id": 28937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004188.wav"
+            ]
+        }
+    },
+    {
+        "id": 28940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008823.wav"
+            ]
+        }
+    },
+    {
+        "id": 28941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001107.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001107.wav"
+            ]
+        }
+    },
+    {
+        "id": 28942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005895.wav"
+            ]
+        }
+    },
+    {
+        "id": 28949,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005160.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005160.wav"
+            ]
+        }
+    },
+    {
+        "id": 28953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004946.wav"
+            ]
+        }
+    },
+    {
+        "id": 28965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006962.wav"
+            ]
+        }
+    },
+    {
+        "id": 28968,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000031.wav"
+            ]
+        }
+    },
+    {
+        "id": 28973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there banjo and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006017.wav"
+            ]
+        }
+    },
+    {
+        "id": 28984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005286.wav"
+            ]
+        }
+    },
+    {
+        "id": 28992,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008828.wav"
+            ]
+        }
+    },
+    {
+        "id": 28998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000522.wav"
+            ]
+        }
+    },
+    {
+        "id": 29000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001529.wav"
+            ]
+        }
+    },
+    {
+        "id": 29002,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003668.wav"
+            ]
+        }
+    },
+    {
+        "id": 29006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005668.wav"
+            ]
+        }
+    },
+    {
+        "id": 29019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006233.wav"
+            ]
+        }
+    },
+    {
+        "id": 29034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008956.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008956.wav"
+            ]
+        }
+    },
+    {
+        "id": 29042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002191.wav"
+            ]
+        }
+    },
+    {
+        "id": 29053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004624.wav"
+            ]
+        }
+    },
+    {
+        "id": 29060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000368.wav"
+            ]
+        }
+    },
+    {
+        "id": 29062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006544.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006544.wav"
+            ]
+        }
+    },
+    {
+        "id": 29065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008415.wav"
+            ]
+        }
+    },
+    {
+        "id": 29070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002992.wav"
+            ]
+        }
+    },
+    {
+        "id": 29080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006185.wav"
+            ]
+        }
+    },
+    {
+        "id": 29090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002775.wav"
+            ]
+        }
+    },
+    {
+        "id": 29095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006908.wav"
+            ]
+        }
+    },
+    {
+        "id": 29102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000995.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000995.wav"
+            ]
+        }
+    },
+    {
+        "id": 29111,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003925.wav"
+            ]
+        }
+    },
+    {
+        "id": 29118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008213.wav"
+            ]
+        }
+    },
+    {
+        "id": 29120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003109.wav"
+            ]
+        }
+    },
+    {
+        "id": 29134,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000497.wav"
+            ]
+        }
+    },
+    {
+        "id": 29135,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004589.wav"
+            ]
+        }
+    },
+    {
+        "id": 29141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and flute instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000353.wav"
+            ]
+        }
+    },
+    {
+        "id": 29147,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002164.wav"
+            ]
+        }
+    },
+    {
+        "id": 29149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000793.wav"
+            ]
+        }
+    },
+    {
+        "id": 29150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006291.wav"
+            ]
+        }
+    },
+    {
+        "id": 29153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006883.wav"
+            ]
+        }
+    },
+    {
+        "id": 29157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004831.wav"
+            ]
+        }
+    },
+    {
+        "id": 29161,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006808.wav"
+            ]
+        }
+    },
+    {
+        "id": 29167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001808.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001808.wav"
+            ]
+        }
+    },
+    {
+        "id": 29172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007537.wav"
+            ]
+        }
+    },
+    {
+        "id": 29177,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007751.wav"
+            ]
+        }
+    },
+    {
+        "id": 29179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005110.wav"
+            ]
+        }
+    },
+    {
+        "id": 29180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005260.wav"
+            ]
+        }
+    },
+    {
+        "id": 29181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007561.wav"
+            ]
+        }
+    },
+    {
+        "id": 29188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008927.wav"
+            ]
+        }
+    },
+    {
+        "id": 29189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007122.wav"
+            ]
+        }
+    },
+    {
+        "id": 29197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006308.wav"
+            ]
+        }
+    },
+    {
+        "id": 29203,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006238.wav"
+            ]
+        }
+    },
+    {
+        "id": 29206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000668.wav"
+            ]
+        }
+    },
+    {
+        "id": 29214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005711.wav"
+            ]
+        }
+    },
+    {
+        "id": 29217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004440.wav"
+            ]
+        }
+    },
+    {
+        "id": 29222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006395.wav"
+            ]
+        }
+    },
+    {
+        "id": 29233,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004507.wav"
+            ]
+        }
+    },
+    {
+        "id": 29243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003718.wav"
+            ]
+        }
+    },
+    {
+        "id": 29244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002475.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002475.wav"
+            ]
+        }
+    },
+    {
+        "id": 29248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005102.wav"
+            ]
+        }
+    },
+    {
+        "id": 29252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001425.wav"
+            ]
+        }
+    },
+    {
+        "id": 29254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001586.wav"
+            ]
+        }
+    },
+    {
+        "id": 29261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006588.wav"
+            ]
+        }
+    },
+    {
+        "id": 29263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002645.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002645.wav"
+            ]
+        }
+    },
+    {
+        "id": 29268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004334.wav"
+            ]
+        }
+    },
+    {
+        "id": 29269,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004745.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004745.wav"
+            ]
+        }
+    },
+    {
+        "id": 29277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007668.wav"
+            ]
+        }
+    },
+    {
+        "id": 29280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001984.wav"
+            ]
+        }
+    },
+    {
+        "id": 29282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005156.wav"
+            ]
+        }
+    },
+    {
+        "id": 29287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002824.wav"
+            ]
+        }
+    },
+    {
+        "id": 29292,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005350.wav"
+            ]
+        }
+    },
+    {
+        "id": 29294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004620.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004620.wav"
+            ]
+        }
+    },
+    {
+        "id": 29297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004596.wav"
+            ]
+        }
+    },
+    {
+        "id": 29306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000451.wav"
+            ]
+        }
+    },
+    {
+        "id": 29307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001791.wav"
+            ]
+        }
+    },
+    {
+        "id": 29315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004000.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004000.wav"
+            ]
+        }
+    },
+    {
+        "id": 29322,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003102.wav"
+            ]
+        }
+    },
+    {
+        "id": 29324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006584.wav"
+            ]
+        }
+    },
+    {
+        "id": 29331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005881.wav"
+            ]
+        }
+    },
+    {
+        "id": 29335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005003.wav"
+            ]
+        }
+    },
+    {
+        "id": 29339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006195.wav"
+            ]
+        }
+    },
+    {
+        "id": 29345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006095.wav"
+            ]
+        }
+    },
+    {
+        "id": 29348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003707.wav"
+            ]
+        }
+    },
+    {
+        "id": 29356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002758.wav"
+            ]
+        }
+    },
+    {
+        "id": 29357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008681.wav"
+            ]
+        }
+    },
+    {
+        "id": 29371,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there xylophone and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005494.wav"
+            ]
+        }
+    },
+    {
+        "id": 29376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the clarinet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005451.wav"
+            ]
+        }
+    },
+    {
+        "id": 29378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and drum instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005987.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005987.wav"
+            ]
+        }
+    },
+    {
+        "id": 29379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000044.wav"
+            ]
+        }
+    },
+    {
+        "id": 29384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003658.wav"
+            ]
+        }
+    },
+    {
+        "id": 29386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there clarinet and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005096.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005096.wav"
+            ]
+        }
+    },
+    {
+        "id": 29392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001984.wav"
+            ]
+        }
+    },
+    {
+        "id": 29397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004403.wav"
+            ]
+        }
+    },
+    {
+        "id": 29398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there saxophone and pipa instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000227.wav"
+            ]
+        }
+    },
+    {
+        "id": 29404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001644.wav"
+            ]
+        }
+    },
+    {
+        "id": 29409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002489.wav"
+            ]
+        }
+    },
+    {
+        "id": 29428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001166.wav"
+            ]
+        }
+    },
+    {
+        "id": 29430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005476.wav"
+            ]
+        }
+    },
+    {
+        "id": 29431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002396.wav"
+            ]
+        }
+    },
+    {
+        "id": 29440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005044.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005044.wav"
+            ]
+        }
+    },
+    {
+        "id": 29446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004391.wav"
+            ]
+        }
+    },
+    {
+        "id": 29447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006779.wav"
+            ]
+        }
+    },
+    {
+        "id": 29452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008018.wav"
+            ]
+        }
+    },
+    {
+        "id": 29454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006700.wav"
+            ]
+        }
+    },
+    {
+        "id": 29456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there guzheng and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007792.wav"
+            ]
+        }
+    },
+    {
+        "id": 29457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007043.wav"
+            ]
+        }
+    },
+    {
+        "id": 29463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bagpipe and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000525.wav"
+            ]
+        }
+    },
+    {
+        "id": 29464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008080.wav"
+            ]
+        }
+    },
+    {
+        "id": 29468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001999.wav"
+            ]
+        }
+    },
+    {
+        "id": 29470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007593.wav"
+            ]
+        }
+    },
+    {
+        "id": 29471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008523.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008523.wav"
+            ]
+        }
+    },
+    {
+        "id": 29475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006603.wav"
+            ]
+        }
+    },
+    {
+        "id": 29482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002670.wav"
+            ]
+        }
+    },
+    {
+        "id": 29487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000533.wav"
+            ]
+        }
+    },
+    {
+        "id": 29492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007928.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007928.wav"
+            ]
+        }
+    },
+    {
+        "id": 29493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002660.wav"
+            ]
+        }
+    },
+    {
+        "id": 29495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000083.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000083.wav"
+            ]
+        }
+    },
+    {
+        "id": 29497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001894.wav"
+            ]
+        }
+    },
+    {
+        "id": 29505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003176.wav"
+            ]
+        }
+    },
+    {
+        "id": 29510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002538.wav"
+            ]
+        }
+    },
+    {
+        "id": 29513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001365.wav"
+            ]
+        }
+    },
+    {
+        "id": 29515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001743.wav"
+            ]
+        }
+    },
+    {
+        "id": 29518,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008062.wav"
+            ]
+        }
+    },
+    {
+        "id": 29520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008830.wav"
+            ]
+        }
+    },
+    {
+        "id": 29524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002541.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002541.wav"
+            ]
+        }
+    },
+    {
+        "id": 29526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002369.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002369.wav"
+            ]
+        }
+    },
+    {
+        "id": 29534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008208.wav"
+            ]
+        }
+    },
+    {
+        "id": 29538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004544.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004544.wav"
+            ]
+        }
+    },
+    {
+        "id": 29549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002401.wav"
+            ]
+        }
+    },
+    {
+        "id": 29559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006409.wav"
+            ]
+        }
+    },
+    {
+        "id": 29566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001925.wav"
+            ]
+        }
+    },
+    {
+        "id": 29569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005714.wav"
+            ]
+        }
+    },
+    {
+        "id": 29573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005182.wav"
+            ]
+        }
+    },
+    {
+        "id": 29576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007670.wav"
+            ]
+        }
+    },
+    {
+        "id": 29583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005164.wav"
+            ]
+        }
+    },
+    {
+        "id": 29585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000689.wav"
+            ]
+        }
+    },
+    {
+        "id": 29587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003767.wav"
+            ]
+        }
+    },
+    {
+        "id": 29589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006879.wav"
+            ]
+        }
+    },
+    {
+        "id": 29593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006412.wav"
+            ]
+        }
+    },
+    {
+        "id": 29595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002320.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002320.wav"
+            ]
+        }
+    },
+    {
+        "id": 29596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and acoustic_guitar instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006331.wav"
+            ]
+        }
+    },
+    {
+        "id": 29597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003754.wav"
+            ]
+        }
+    },
+    {
+        "id": 29599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002624.wav"
+            ]
+        }
+    },
+    {
+        "id": 29600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005526.wav"
+            ]
+        }
+    },
+    {
+        "id": 29601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008418.wav"
+            ]
+        }
+    },
+    {
+        "id": 29618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005802.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005802.wav"
+            ]
+        }
+    },
+    {
+        "id": 29621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004132.wav"
+            ]
+        }
+    },
+    {
+        "id": 29624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008904.wav"
+            ]
+        }
+    },
+    {
+        "id": 29628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003361.wav"
+            ]
+        }
+    },
+    {
+        "id": 29632,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002599.wav"
+            ]
+        }
+    },
+    {
+        "id": 29637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002466.wav"
+            ]
+        }
+    },
+    {
+        "id": 29646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003885.wav"
+            ]
+        }
+    },
+    {
+        "id": 29668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006108.wav"
+            ]
+        }
+    },
+    {
+        "id": 29670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006896.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006896.wav"
+            ]
+        }
+    },
+    {
+        "id": 29678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and drum instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004305.wav"
+            ]
+        }
+    },
+    {
+        "id": 29683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000628.wav"
+            ]
+        }
+    },
+    {
+        "id": 29684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004324.wav"
+            ]
+        }
+    },
+    {
+        "id": 29689,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006791.wav"
+            ]
+        }
+    },
+    {
+        "id": 29692,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004985.wav"
+            ]
+        }
+    },
+    {
+        "id": 29709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich tuba makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001756.wav"
+            ]
+        }
+    },
+    {
+        "id": 29716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002918.wav"
+            ]
+        }
+    },
+    {
+        "id": 29718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich acoustic_guitar makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001652.wav"
+            ]
+        }
+    },
+    {
+        "id": 29719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005295.wav"
+            ]
+        }
+    },
+    {
+        "id": 29720,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006810.wav"
+            ]
+        }
+    },
+    {
+        "id": 29721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008581.wav"
+            ]
+        }
+    },
+    {
+        "id": 29724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002584.wav"
+            ]
+        }
+    },
+    {
+        "id": 29737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000531.wav"
+            ]
+        }
+    },
+    {
+        "id": 29742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008398.wav"
+            ]
+        }
+    },
+    {
+        "id": 29746,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and flute instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005128.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005128.wav"
+            ]
+        }
+    },
+    {
+        "id": 29749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008349.wav"
+            ]
+        }
+    },
+    {
+        "id": 29751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003688.wav"
+            ]
+        }
+    },
+    {
+        "id": 29756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004732.wav"
+            ]
+        }
+    },
+    {
+        "id": 29759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000777.wav"
+            ]
+        }
+    },
+    {
+        "id": 29766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000598.wav"
+            ]
+        }
+    },
+    {
+        "id": 29772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005144.wav"
+            ]
+        }
+    },
+    {
+        "id": 29776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there cello and drum instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007027.wav"
+            ]
+        }
+    },
+    {
+        "id": 29781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008589.wav"
+            ]
+        }
+    },
+    {
+        "id": 29786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004339.wav"
+            ]
+        }
+    },
+    {
+        "id": 29810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there acoustic_guitar and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006848.wav"
+            ]
+        }
+    },
+    {
+        "id": 29812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007566.wav"
+            ]
+        }
+    },
+    {
+        "id": 29814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004477.wav"
+            ]
+        }
+    },
+    {
+        "id": 29819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000914.wav"
+            ]
+        }
+    },
+    {
+        "id": 29821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002820.wav"
+            ]
+        }
+    },
+    {
+        "id": 29825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006676.wav"
+            ]
+        }
+    },
+    {
+        "id": 29832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001772.wav"
+            ]
+        }
+    },
+    {
+        "id": 29838,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002653.wav"
+            ]
+        }
+    },
+    {
+        "id": 29842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008899.wav"
+            ]
+        }
+    },
+    {
+        "id": 29845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001390.wav"
+            ]
+        }
+    },
+    {
+        "id": 29849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006449.wav"
+            ]
+        }
+    },
+    {
+        "id": 29860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001322.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001322.wav"
+            ]
+        }
+    },
+    {
+        "id": 29867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008694.wav"
+            ]
+        }
+    },
+    {
+        "id": 29868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002944.wav"
+            ]
+        }
+    },
+    {
+        "id": 29884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004492.wav"
+            ]
+        }
+    },
+    {
+        "id": 29886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000103.wav"
+            ]
+        }
+    },
+    {
+        "id": 29891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001912.wav"
+            ]
+        }
+    },
+    {
+        "id": 29892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004540.wav"
+            ]
+        }
+    },
+    {
+        "id": 29893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003141.wav"
+            ]
+        }
+    },
+    {
+        "id": 29894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006175.wav"
+            ]
+        }
+    },
+    {
+        "id": 29896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007536.wav"
+            ]
+        }
+    },
+    {
+        "id": 29907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004677.wav"
+            ]
+        }
+    },
+    {
+        "id": 29913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000606.wav"
+            ]
+        }
+    },
+    {
+        "id": 29915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001721.wav"
+            ]
+        }
+    },
+    {
+        "id": 29921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005853.wav"
+            ]
+        }
+    },
+    {
+        "id": 29922,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003270.wav"
+            ]
+        }
+    },
+    {
+        "id": 29931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003430.wav"
+            ]
+        }
+    },
+    {
+        "id": 29936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and ukulele instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003770.wav"
+            ]
+        }
+    },
+    {
+        "id": 29951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003313.wav"
+            ]
+        }
+    },
+    {
+        "id": 29957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007719.wav"
+            ]
+        }
+    },
+    {
+        "id": 29958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004518.wav"
+            ]
+        }
+    },
+    {
+        "id": 29964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000274.wav"
+            ]
+        }
+    },
+    {
+        "id": 29966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002275.wav"
+            ]
+        }
+    },
+    {
+        "id": 29973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001139.wav"
+            ]
+        }
+    },
+    {
+        "id": 29979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the middle more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006968.wav"
+            ]
+        }
+    },
+    {
+        "id": 29980,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001207.wav"
+            ]
+        }
+    },
+    {
+        "id": 29985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006981.wav"
+            ]
+        }
+    },
+    {
+        "id": 29996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000634.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000634.wav"
+            ]
+        }
+    },
+    {
+        "id": 29997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006387.wav"
+            ]
+        }
+    },
+    {
+        "id": 29999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003074.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003074.wav"
+            ]
+        }
+    },
+    {
+        "id": 30008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003547.wav"
+            ]
+        }
+    },
+    {
+        "id": 30014,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004521.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004521.wav"
+            ]
+        }
+    },
+    {
+        "id": 30015,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008560.wav"
+            ]
+        }
+    },
+    {
+        "id": 30023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and cello instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005200.wav"
+            ]
+        }
+    },
+    {
+        "id": 30025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003467.wav"
+            ]
+        }
+    },
+    {
+        "id": 30032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000952.wav"
+            ]
+        }
+    },
+    {
+        "id": 30033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and drum instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004057.wav"
+            ]
+        }
+    },
+    {
+        "id": 30034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004663.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004663.wav"
+            ]
+        }
+    },
+    {
+        "id": 30040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004935.wav"
+            ]
+        }
+    },
+    {
+        "id": 30045,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000649.wav"
+            ]
+        }
+    },
+    {
+        "id": 30046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001625.wav"
+            ]
+        }
+    },
+    {
+        "id": 30049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there erhu and guzheng instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005602.wav"
+            ]
+        }
+    },
+    {
+        "id": 30051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007266.wav"
+            ]
+        }
+    },
+    {
+        "id": 30054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003661.wav"
+            ]
+        }
+    },
+    {
+        "id": 30055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007789.wav"
+            ]
+        }
+    },
+    {
+        "id": 30062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001945.wav"
+            ]
+        }
+    },
+    {
+        "id": 30066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005655.wav"
+            ]
+        }
+    },
+    {
+        "id": 30069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005930.wav"
+            ]
+        }
+    },
+    {
+        "id": 30072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007621.wav"
+            ]
+        }
+    },
+    {
+        "id": 30074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000404.wav"
+            ]
+        }
+    },
+    {
+        "id": 30075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008643.wav"
+            ]
+        }
+    },
+    {
+        "id": 30078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004827.wav"
+            ]
+        }
+    },
+    {
+        "id": 30079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001061.wav"
+            ]
+        }
+    },
+    {
+        "id": 30087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003780.wav"
+            ]
+        }
+    },
+    {
+        "id": 30089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004027.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004027.wav"
+            ]
+        }
+    },
+    {
+        "id": 30090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005700.wav"
+            ]
+        }
+    },
+    {
+        "id": 30092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004892.wav"
+            ]
+        }
+    },
+    {
+        "id": 30093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008059.wav"
+            ]
+        }
+    },
+    {
+        "id": 30097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000644.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000644.wav"
+            ]
+        }
+    },
+    {
+        "id": 30100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005209.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005209.wav"
+            ]
+        }
+    },
+    {
+        "id": 30119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and erhu instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008536.wav"
+            ]
+        }
+    },
+    {
+        "id": 30128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there flute and accordion instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000068.wav"
+            ]
+        }
+    },
+    {
+        "id": 30129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002566.wav"
+            ]
+        }
+    },
+    {
+        "id": 30137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there accordion and violin instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008908.wav"
+            ]
+        }
+    },
+    {
+        "id": 30143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006245.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006245.wav"
+            ]
+        }
+    },
+    {
+        "id": 30151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there electric_bass and saxophone instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008012.wav"
+            ]
+        }
+    },
+    {
+        "id": 30153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003126.wav"
+            ]
+        }
+    },
+    {
+        "id": 30157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004157.wav"
+            ]
+        }
+    },
+    {
+        "id": 30167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000824.wav"
+            ]
+        }
+    },
+    {
+        "id": 30173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there violin and banjo instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007059.wav"
+            ]
+        }
+    },
+    {
+        "id": 30180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007196.wav"
+            ]
+        }
+    },
+    {
+        "id": 30182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002382.wav"
+            ]
+        }
+    },
+    {
+        "id": 30184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004760.wav"
+            ]
+        }
+    },
+    {
+        "id": 30186,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005195.wav"
+            ]
+        }
+    },
+    {
+        "id": 30191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008911.wav"
+            ]
+        }
+    },
+    {
+        "id": 30192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there ukulele and tuba instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001374.wav"
+            ]
+        }
+    },
+    {
+        "id": 30202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002697.wav"
+            ]
+        }
+    },
+    {
+        "id": 30207,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003157.wav"
+            ]
+        }
+    },
+    {
+        "id": 30220,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006802.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006802.wav"
+            ]
+        }
+    },
+    {
+        "id": 30225,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002992.wav"
+            ]
+        }
+    },
+    {
+        "id": 30231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there piano and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000835.wav"
+            ]
+        }
+    },
+    {
+        "id": 30246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there bassoon and clarinet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007628.wav"
+            ]
+        }
+    },
+    {
+        "id": 30250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the performance?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003442.wav"
+            ]
+        }
+    },
+    {
+        "id": 30255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001399.wav"
+            ]
+        }
+    },
+    {
+        "id": 30258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the drum on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007538.wav"
+            ]
+        }
+    },
+    {
+        "id": 30261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there tuba and piano instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000919.wav"
+            ]
+        }
+    },
+    {
+        "id": 30262,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nAre there trumpet and trumpet instruments in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002360.wav"
+            ]
+        }
+    },
+    {
+        "id": 30271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000798.wav"
+            ]
+        }
+    },
+    {
+        "id": 30279,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002785.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002785.wav"
+            ]
+        }
+    },
+    {
+        "id": 30285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005427.wav"
+            ]
+        }
+    },
+    {
+        "id": 30298,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007972.wav"
+            ]
+        }
+    },
+    {
+        "id": 30304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003532.wav"
+            ]
+        }
+    },
+    {
+        "id": 30308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007737.wav"
+            ]
+        }
+    },
+    {
+        "id": 30322,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004452.wav"
+            ]
+        }
+    },
+    {
+        "id": 30324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005521.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005521.wav"
+            ]
+        }
+    },
+    {
+        "id": 30327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000670.wav"
+            ]
+        }
+    },
+    {
+        "id": 30331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000698.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000698.wav"
+            ]
+        }
+    },
+    {
+        "id": 30342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004886.wav"
+            ]
+        }
+    },
+    {
+        "id": 30353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007836.wav"
+            ]
+        }
+    },
+    {
+        "id": 30354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000653.wav"
+            ]
+        }
+    },
+    {
+        "id": 30359,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000675.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000675.wav"
+            ]
+        }
+    },
+    {
+        "id": 30365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008875.wav"
+            ]
+        }
+    },
+    {
+        "id": 30372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005672.wav"
+            ]
+        }
+    },
+    {
+        "id": 30374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007934.wav"
+            ]
+        }
+    },
+    {
+        "id": 30379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008071.wav"
+            ]
+        }
+    },
+    {
+        "id": 30381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000774.wav"
+            ]
+        }
+    },
+    {
+        "id": 30386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001699.wav"
+            ]
+        }
+    },
+    {
+        "id": 30389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003515.wav"
+            ]
+        }
+    },
+    {
+        "id": 30401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006091.wav"
+            ]
+        }
+    },
+    {
+        "id": 30406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006072.wav"
+            ]
+        }
+    },
+    {
+        "id": 30408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003530.wav"
+            ]
+        }
+    },
+    {
+        "id": 30409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002415.wav"
+            ]
+        }
+    },
+    {
+        "id": 30411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002669.wav"
+            ]
+        }
+    },
+    {
+        "id": 30412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003081.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003081.wav"
+            ]
+        }
+    },
+    {
+        "id": 30417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006582.wav"
+            ]
+        }
+    },
+    {
+        "id": 30418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001278.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001278.wav"
+            ]
+        }
+    },
+    {
+        "id": 30427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008522.wav"
+            ]
+        }
+    },
+    {
+        "id": 30440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000091.wav"
+            ]
+        }
+    },
+    {
+        "id": 30441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002300.wav"
+            ]
+        }
+    },
+    {
+        "id": 30454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007328.wav"
+            ]
+        }
+    },
+    {
+        "id": 30455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002143.wav"
+            ]
+        }
+    },
+    {
+        "id": 30473,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005637.wav"
+            ]
+        }
+    },
+    {
+        "id": 30476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003593.wav"
+            ]
+        }
+    },
+    {
+        "id": 30483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001029.wav"
+            ]
+        }
+    },
+    {
+        "id": 30486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich acoustic_guitar makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002477.wav"
+            ]
+        }
+    },
+    {
+        "id": 30490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007416.wav"
+            ]
+        }
+    },
+    {
+        "id": 30496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001249.wav"
+            ]
+        }
+    },
+    {
+        "id": 30505,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006740.wav"
+            ]
+        }
+    },
+    {
+        "id": 30511,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002495.wav"
+            ]
+        }
+    },
+    {
+        "id": 30512,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005172.wav"
+            ]
+        }
+    },
+    {
+        "id": 30533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003020.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003020.wav"
+            ]
+        }
+    },
+    {
+        "id": 30534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002743.wav"
+            ]
+        }
+    },
+    {
+        "id": 30550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich xylophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003650.wav"
+            ]
+        }
+    },
+    {
+        "id": 30553,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004869.wav"
+            ]
+        }
+    },
+    {
+        "id": 30557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008493.wav"
+            ]
+        }
+    },
+    {
+        "id": 30558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003053.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003053.wav"
+            ]
+        }
+    },
+    {
+        "id": 30560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001311.wav"
+            ]
+        }
+    },
+    {
+        "id": 30561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008143.wav"
+            ]
+        }
+    },
+    {
+        "id": 30579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000908.wav"
+            ]
+        }
+    },
+    {
+        "id": 30583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007366.wav"
+            ]
+        }
+    },
+    {
+        "id": 30584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006124.wav"
+            ]
+        }
+    },
+    {
+        "id": 30589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002297.wav"
+            ]
+        }
+    },
+    {
+        "id": 30593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002421.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002421.wav"
+            ]
+        }
+    },
+    {
+        "id": 30596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008979.wav"
+            ]
+        }
+    },
+    {
+        "id": 30608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007446.wav"
+            ]
+        }
+    },
+    {
+        "id": 30609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005926.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005926.wav"
+            ]
+        }
+    },
+    {
+        "id": 30611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001619.wav"
+            ]
+        }
+    },
+    {
+        "id": 30612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007486.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007486.wav"
+            ]
+        }
+    },
+    {
+        "id": 30614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007318.wav"
+            ]
+        }
+    },
+    {
+        "id": 30618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007586.wav"
+            ]
+        }
+    },
+    {
+        "id": 30629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007729.wav"
+            ]
+        }
+    },
+    {
+        "id": 30633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004570.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004570.wav"
+            ]
+        }
+    },
+    {
+        "id": 30637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005941.wav"
+            ]
+        }
+    },
+    {
+        "id": 30640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008382.wav"
+            ]
+        }
+    },
+    {
+        "id": 30652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002805.wav"
+            ]
+        }
+    },
+    {
+        "id": 30659,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006854.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006854.wav"
+            ]
+        }
+    },
+    {
+        "id": 30669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006900.wav"
+            ]
+        }
+    },
+    {
+        "id": 30675,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000515.wav"
+            ]
+        }
+    },
+    {
+        "id": 30680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004296.wav"
+            ]
+        }
+    },
+    {
+        "id": 30694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006101.wav"
+            ]
+        }
+    },
+    {
+        "id": 30695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007851.wav"
+            ]
+        }
+    },
+    {
+        "id": 30699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004831.wav"
+            ]
+        }
+    },
+    {
+        "id": 30706,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007018.wav"
+            ]
+        }
+    },
+    {
+        "id": 30708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006430.wav"
+            ]
+        }
+    },
+    {
+        "id": 30713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008479.wav"
+            ]
+        }
+    },
+    {
+        "id": 30716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005938.wav"
+            ]
+        }
+    },
+    {
+        "id": 30728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008553.wav"
+            ]
+        }
+    },
+    {
+        "id": 30736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002908.wav"
+            ]
+        }
+    },
+    {
+        "id": 30750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003810.wav"
+            ]
+        }
+    },
+    {
+        "id": 30753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002220.wav"
+            ]
+        }
+    },
+    {
+        "id": 30758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000649.wav"
+            ]
+        }
+    },
+    {
+        "id": 30759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007603.wav"
+            ]
+        }
+    },
+    {
+        "id": 30765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004780.wav"
+            ]
+        }
+    },
+    {
+        "id": 30772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007805.wav"
+            ]
+        }
+    },
+    {
+        "id": 30781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003344.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003344.wav"
+            ]
+        }
+    },
+    {
+        "id": 30795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004734.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004734.wav"
+            ]
+        }
+    },
+    {
+        "id": 30800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004251.wav"
+            ]
+        }
+    },
+    {
+        "id": 30819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008522.wav"
+            ]
+        }
+    },
+    {
+        "id": 30821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006000.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006000.wav"
+            ]
+        }
+    },
+    {
+        "id": 30827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002865.wav"
+            ]
+        }
+    },
+    {
+        "id": 30829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005601.wav"
+            ]
+        }
+    },
+    {
+        "id": 30831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003680.wav"
+            ]
+        }
+    },
+    {
+        "id": 30833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001328.wav"
+            ]
+        }
+    },
+    {
+        "id": 30840,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005304.wav"
+            ]
+        }
+    },
+    {
+        "id": 30841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004629.wav"
+            ]
+        }
+    },
+    {
+        "id": 30850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006026.wav"
+            ]
+        }
+    },
+    {
+        "id": 30867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001289.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001289.wav"
+            ]
+        }
+    },
+    {
+        "id": 30868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007492.wav"
+            ]
+        }
+    },
+    {
+        "id": 30869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004859.wav"
+            ]
+        }
+    },
+    {
+        "id": 30870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000721.wav"
+            ]
+        }
+    },
+    {
+        "id": 30873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002078.wav"
+            ]
+        }
+    },
+    {
+        "id": 30883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007744.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007744.wav"
+            ]
+        }
+    },
+    {
+        "id": 30885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007129.wav"
+            ]
+        }
+    },
+    {
+        "id": 30894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005937.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005937.wav"
+            ]
+        }
+    },
+    {
+        "id": 30909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007476.wav"
+            ]
+        }
+    },
+    {
+        "id": 30912,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001079.wav"
+            ]
+        }
+    },
+    {
+        "id": 30913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002379.wav"
+            ]
+        }
+    },
+    {
+        "id": 30916,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008773.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008773.wav"
+            ]
+        }
+    },
+    {
+        "id": 30920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008042.wav"
+            ]
+        }
+    },
+    {
+        "id": 30934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003238.wav"
+            ]
+        }
+    },
+    {
+        "id": 30935,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006827.wav"
+            ]
+        }
+    },
+    {
+        "id": 30941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000575.wav"
+            ]
+        }
+    },
+    {
+        "id": 30957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002626.wav"
+            ]
+        }
+    },
+    {
+        "id": 30970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004198.wav"
+            ]
+        }
+    },
+    {
+        "id": 30982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004721.wav"
+            ]
+        }
+    },
+    {
+        "id": 30983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006947.wav"
+            ]
+        }
+    },
+    {
+        "id": 30984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000501.wav"
+            ]
+        }
+    },
+    {
+        "id": 30990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba playing longer than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002988.wav"
+            ]
+        }
+    },
+    {
+        "id": 30993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004996.wav"
+            ]
+        }
+    },
+    {
+        "id": 30997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004106.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004106.wav"
+            ]
+        }
+    },
+    {
+        "id": 31005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007923.wav"
+            ]
+        }
+    },
+    {
+        "id": 31012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005934.wav"
+            ]
+        }
+    },
+    {
+        "id": 31015,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001592.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001592.wav"
+            ]
+        }
+    },
+    {
+        "id": 31019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003430.wav"
+            ]
+        }
+    },
+    {
+        "id": 31030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000966.wav"
+            ]
+        }
+    },
+    {
+        "id": 31034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000401.wav"
+            ]
+        }
+    },
+    {
+        "id": 31038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006471.wav"
+            ]
+        }
+    },
+    {
+        "id": 31048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005578.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005578.wav"
+            ]
+        }
+    },
+    {
+        "id": 31049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007424.wav"
+            ]
+        }
+    },
+    {
+        "id": 31050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004995.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004995.wav"
+            ]
+        }
+    },
+    {
+        "id": 31053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007800.wav"
+            ]
+        }
+    },
+    {
+        "id": 31056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004144.wav"
+            ]
+        }
+    },
+    {
+        "id": 31058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001593.wav"
+            ]
+        }
+    },
+    {
+        "id": 31069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005612.wav"
+            ]
+        }
+    },
+    {
+        "id": 31078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006227.wav"
+            ]
+        }
+    },
+    {
+        "id": 31080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008946.wav"
+            ]
+        }
+    },
+    {
+        "id": 31084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008254.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008254.wav"
+            ]
+        }
+    },
+    {
+        "id": 31089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008818.wav"
+            ]
+        }
+    },
+    {
+        "id": 31091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001395.wav"
+            ]
+        }
+    },
+    {
+        "id": 31100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002409.wav"
+            ]
+        }
+    },
+    {
+        "id": 31102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008378.wav"
+            ]
+        }
+    },
+    {
+        "id": 31105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000271.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000271.wav"
+            ]
+        }
+    },
+    {
+        "id": 31113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008087.wav"
+            ]
+        }
+    },
+    {
+        "id": 31114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005360.wav"
+            ]
+        }
+    },
+    {
+        "id": 31116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007781.wav"
+            ]
+        }
+    },
+    {
+        "id": 31125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004467.wav"
+            ]
+        }
+    },
+    {
+        "id": 31132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006757.wav"
+            ]
+        }
+    },
+    {
+        "id": 31133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001938.wav"
+            ]
+        }
+    },
+    {
+        "id": 31134,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003330.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003330.wav"
+            ]
+        }
+    },
+    {
+        "id": 31142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000219.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000219.wav"
+            ]
+        }
+    },
+    {
+        "id": 31153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005186.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005186.wav"
+            ]
+        }
+    },
+    {
+        "id": 31155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004419.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004419.wav"
+            ]
+        }
+    },
+    {
+        "id": 31161,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006043.wav"
+            ]
+        }
+    },
+    {
+        "id": 31165,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008461.wav"
+            ]
+        }
+    },
+    {
+        "id": 31166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002948.wav"
+            ]
+        }
+    },
+    {
+        "id": 31174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006359.wav"
+            ]
+        }
+    },
+    {
+        "id": 31179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004004.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004004.wav"
+            ]
+        }
+    },
+    {
+        "id": 31186,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006097.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006097.wav"
+            ]
+        }
+    },
+    {
+        "id": 31189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000391.wav"
+            ]
+        }
+    },
+    {
+        "id": 31194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002386.wav"
+            ]
+        }
+    },
+    {
+        "id": 31200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000573.wav"
+            ]
+        }
+    },
+    {
+        "id": 31206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005585.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005585.wav"
+            ]
+        }
+    },
+    {
+        "id": 31209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005353.wav"
+            ]
+        }
+    },
+    {
+        "id": 31215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich tuba makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004026.wav"
+            ]
+        }
+    },
+    {
+        "id": 31224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004714.wav"
+            ]
+        }
+    },
+    {
+        "id": 31227,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006198.wav"
+            ]
+        }
+    },
+    {
+        "id": 31229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004350.wav"
+            ]
+        }
+    },
+    {
+        "id": 31230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008089.wav"
+            ]
+        }
+    },
+    {
+        "id": 31232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002622.wav"
+            ]
+        }
+    },
+    {
+        "id": 31234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005730.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005730.wav"
+            ]
+        }
+    },
+    {
+        "id": 31238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000293.wav"
+            ]
+        }
+    },
+    {
+        "id": 31247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003306.wav"
+            ]
+        }
+    },
+    {
+        "id": 31251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000769.wav"
+            ]
+        }
+    },
+    {
+        "id": 31253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006492.wav"
+            ]
+        }
+    },
+    {
+        "id": 31256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000327.wav"
+            ]
+        }
+    },
+    {
+        "id": 31257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006150.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006150.wav"
+            ]
+        }
+    },
+    {
+        "id": 31262,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001390.wav"
+            ]
+        }
+    },
+    {
+        "id": 31263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005877.wav"
+            ]
+        }
+    },
+    {
+        "id": 31273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008877.wav"
+            ]
+        }
+    },
+    {
+        "id": 31275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007643.wav"
+            ]
+        }
+    },
+    {
+        "id": 31277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003748.wav"
+            ]
+        }
+    },
+    {
+        "id": 31287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007984.wav"
+            ]
+        }
+    },
+    {
+        "id": 31291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007307.wav"
+            ]
+        }
+    },
+    {
+        "id": 31293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007179.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007179.wav"
+            ]
+        }
+    },
+    {
+        "id": 31294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003378.wav"
+            ]
+        }
+    },
+    {
+        "id": 31296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003865.wav"
+            ]
+        }
+    },
+    {
+        "id": 31297,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007451.wav"
+            ]
+        }
+    },
+    {
+        "id": 31301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000629.wav"
+            ]
+        }
+    },
+    {
+        "id": 31313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008028.wav"
+            ]
+        }
+    },
+    {
+        "id": 31319,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004642.wav"
+            ]
+        }
+    },
+    {
+        "id": 31320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005425.wav"
+            ]
+        }
+    },
+    {
+        "id": 31326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002082.wav"
+            ]
+        }
+    },
+    {
+        "id": 31331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003385.wav"
+            ]
+        }
+    },
+    {
+        "id": 31342,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005627.wav"
+            ]
+        }
+    },
+    {
+        "id": 31343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000428.wav"
+            ]
+        }
+    },
+    {
+        "id": 31346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004191.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004191.wav"
+            ]
+        }
+    },
+    {
+        "id": 31347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001346.wav"
+            ]
+        }
+    },
+    {
+        "id": 31349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002711.wav"
+            ]
+        }
+    },
+    {
+        "id": 31357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000367.wav"
+            ]
+        }
+    },
+    {
+        "id": 31359,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006092.wav"
+            ]
+        }
+    },
+    {
+        "id": 31361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004924.wav"
+            ]
+        }
+    },
+    {
+        "id": 31362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007260.wav"
+            ]
+        }
+    },
+    {
+        "id": 31366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005522.wav"
+            ]
+        }
+    },
+    {
+        "id": 31367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005548.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005548.wav"
+            ]
+        }
+    },
+    {
+        "id": 31372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008517.wav"
+            ]
+        }
+    },
+    {
+        "id": 31374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005919.wav"
+            ]
+        }
+    },
+    {
+        "id": 31382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004538.wav"
+            ]
+        }
+    },
+    {
+        "id": 31389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006111.wav"
+            ]
+        }
+    },
+    {
+        "id": 31390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004242.wav"
+            ]
+        }
+    },
+    {
+        "id": 31393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006540.wav"
+            ]
+        }
+    },
+    {
+        "id": 31399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007746.wav"
+            ]
+        }
+    },
+    {
+        "id": 31401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007199.wav"
+            ]
+        }
+    },
+    {
+        "id": 31408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007171.wav"
+            ]
+        }
+    },
+    {
+        "id": 31409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003735.wav"
+            ]
+        }
+    },
+    {
+        "id": 31410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001863.wav"
+            ]
+        }
+    },
+    {
+        "id": 31412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005514.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005514.wav"
+            ]
+        }
+    },
+    {
+        "id": 31414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007921.wav"
+            ]
+        }
+    },
+    {
+        "id": 31416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003356.wav"
+            ]
+        }
+    },
+    {
+        "id": 31426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002743.wav"
+            ]
+        }
+    },
+    {
+        "id": 31427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003612.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003612.wav"
+            ]
+        }
+    },
+    {
+        "id": 31430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003274.wav"
+            ]
+        }
+    },
+    {
+        "id": 31436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003678.wav"
+            ]
+        }
+    },
+    {
+        "id": 31438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008215.wav"
+            ]
+        }
+    },
+    {
+        "id": 31440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001061.wav"
+            ]
+        }
+    },
+    {
+        "id": 31443,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008798.wav"
+            ]
+        }
+    },
+    {
+        "id": 31448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008535.wav"
+            ]
+        }
+    },
+    {
+        "id": 31457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003075.wav"
+            ]
+        }
+    },
+    {
+        "id": 31463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002688.wav"
+            ]
+        }
+    },
+    {
+        "id": 31465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000154.wav"
+            ]
+        }
+    },
+    {
+        "id": 31467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002272.wav"
+            ]
+        }
+    },
+    {
+        "id": 31475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008933.wav"
+            ]
+        }
+    },
+    {
+        "id": 31487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003815.wav"
+            ]
+        }
+    },
+    {
+        "id": 31490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass on the left more rhythmic than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007853.wav"
+            ]
+        }
+    },
+    {
+        "id": 31497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007223.wav"
+            ]
+        }
+    },
+    {
+        "id": 31499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003333.wav"
+            ]
+        }
+    },
+    {
+        "id": 31502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008136.wav"
+            ]
+        }
+    },
+    {
+        "id": 31504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003809.wav"
+            ]
+        }
+    },
+    {
+        "id": 31512,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002135.wav"
+            ]
+        }
+    },
+    {
+        "id": 31520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002745.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002745.wav"
+            ]
+        }
+    },
+    {
+        "id": 31526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003873.wav"
+            ]
+        }
+    },
+    {
+        "id": 31531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005994.wav"
+            ]
+        }
+    },
+    {
+        "id": 31534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003258.wav"
+            ]
+        }
+    },
+    {
+        "id": 31539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005814.wav"
+            ]
+        }
+    },
+    {
+        "id": 31541,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006260.wav"
+            ]
+        }
+    },
+    {
+        "id": 31547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004433.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004433.wav"
+            ]
+        }
+    },
+    {
+        "id": 31553,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001374.wav"
+            ]
+        }
+    },
+    {
+        "id": 31560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003555.wav"
+            ]
+        }
+    },
+    {
+        "id": 31568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000910.wav"
+            ]
+        }
+    },
+    {
+        "id": 31569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007072.wav"
+            ]
+        }
+    },
+    {
+        "id": 31575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006763.wav"
+            ]
+        }
+    },
+    {
+        "id": 31576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002286.wav"
+            ]
+        }
+    },
+    {
+        "id": 31578,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004518.wav"
+            ]
+        }
+    },
+    {
+        "id": 31581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000403.wav"
+            ]
+        }
+    },
+    {
+        "id": 31582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005764.wav"
+            ]
+        }
+    },
+    {
+        "id": 31588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008133.wav"
+            ]
+        }
+    },
+    {
+        "id": 31589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001054.wav"
+            ]
+        }
+    },
+    {
+        "id": 31590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000649.wav"
+            ]
+        }
+    },
+    {
+        "id": 31593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006635.wav"
+            ]
+        }
+    },
+    {
+        "id": 31601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008650.wav"
+            ]
+        }
+    },
+    {
+        "id": 31604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007211.wav"
+            ]
+        }
+    },
+    {
+        "id": 31612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008794.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008794.wav"
+            ]
+        }
+    },
+    {
+        "id": 31613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007217.wav"
+            ]
+        }
+    },
+    {
+        "id": 31617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008850.wav"
+            ]
+        }
+    },
+    {
+        "id": 31620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000630.wav"
+            ]
+        }
+    },
+    {
+        "id": 31623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002488.wav"
+            ]
+        }
+    },
+    {
+        "id": 31624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003663.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003663.wav"
+            ]
+        }
+    },
+    {
+        "id": 31626,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007775.wav"
+            ]
+        }
+    },
+    {
+        "id": 31631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004984.wav"
+            ]
+        }
+    },
+    {
+        "id": 31633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007750.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007750.wav"
+            ]
+        }
+    },
+    {
+        "id": 31635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008110.wav"
+            ]
+        }
+    },
+    {
+        "id": 31637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007426.wav"
+            ]
+        }
+    },
+    {
+        "id": 31645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006917.wav"
+            ]
+        }
+    },
+    {
+        "id": 31648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007934.wav"
+            ]
+        }
+    },
+    {
+        "id": 31649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008259.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008259.wav"
+            ]
+        }
+    },
+    {
+        "id": 31651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007859.wav"
+            ]
+        }
+    },
+    {
+        "id": 31656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002306.wav"
+            ]
+        }
+    },
+    {
+        "id": 31668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003743.wav"
+            ]
+        }
+    },
+    {
+        "id": 31673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003887.wav"
+            ]
+        }
+    },
+    {
+        "id": 31691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007979.wav"
+            ]
+        }
+    },
+    {
+        "id": 31692,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the clarinet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003272.wav"
+            ]
+        }
+    },
+    {
+        "id": 31698,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003456.wav"
+            ]
+        }
+    },
+    {
+        "id": 31701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005206.wav"
+            ]
+        }
+    },
+    {
+        "id": 31704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007974.wav"
+            ]
+        }
+    },
+    {
+        "id": 31712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004777.wav"
+            ]
+        }
+    },
+    {
+        "id": 31718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007473.wav"
+            ]
+        }
+    },
+    {
+        "id": 31720,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003339.wav"
+            ]
+        }
+    },
+    {
+        "id": 31724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008599.wav"
+            ]
+        }
+    },
+    {
+        "id": 31729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007280.wav"
+            ]
+        }
+    },
+    {
+        "id": 31742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002966.wav"
+            ]
+        }
+    },
+    {
+        "id": 31745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003894.wav"
+            ]
+        }
+    },
+    {
+        "id": 31749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004438.wav"
+            ]
+        }
+    },
+    {
+        "id": 31756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007054.wav"
+            ]
+        }
+    },
+    {
+        "id": 31760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006550.wav"
+            ]
+        }
+    },
+    {
+        "id": 31761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000203.wav"
+            ]
+        }
+    },
+    {
+        "id": 31771,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001224.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001224.wav"
+            ]
+        }
+    },
+    {
+        "id": 31773,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002431.wav"
+            ]
+        }
+    },
+    {
+        "id": 31779,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008848.wav"
+            ]
+        }
+    },
+    {
+        "id": 31780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002038.wav"
+            ]
+        }
+    },
+    {
+        "id": 31788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002705.wav"
+            ]
+        }
+    },
+    {
+        "id": 31790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007127.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007127.wav"
+            ]
+        }
+    },
+    {
+        "id": 31793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006370.wav"
+            ]
+        }
+    },
+    {
+        "id": 31797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005664.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005664.wav"
+            ]
+        }
+    },
+    {
+        "id": 31798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007257.wav"
+            ]
+        }
+    },
+    {
+        "id": 31799,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003429.wav"
+            ]
+        }
+    },
+    {
+        "id": 31804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000356.wav"
+            ]
+        }
+    },
+    {
+        "id": 31810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007367.wav"
+            ]
+        }
+    },
+    {
+        "id": 31813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006203.wav"
+            ]
+        }
+    },
+    {
+        "id": 31816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004174.wav"
+            ]
+        }
+    },
+    {
+        "id": 31823,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002792.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002792.wav"
+            ]
+        }
+    },
+    {
+        "id": 31832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002642.wav"
+            ]
+        }
+    },
+    {
+        "id": 31840,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003299.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003299.wav"
+            ]
+        }
+    },
+    {
+        "id": 31844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006891.wav"
+            ]
+        }
+    },
+    {
+        "id": 31845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003437.wav"
+            ]
+        }
+    },
+    {
+        "id": 31856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004282.wav"
+            ]
+        }
+    },
+    {
+        "id": 31858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005103.wav"
+            ]
+        }
+    },
+    {
+        "id": 31864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004955.wav"
+            ]
+        }
+    },
+    {
+        "id": 31877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003972.wav"
+            ]
+        }
+    },
+    {
+        "id": 31882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000445.wav"
+            ]
+        }
+    },
+    {
+        "id": 31890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003081.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003081.wav"
+            ]
+        }
+    },
+    {
+        "id": 31892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002038.wav"
+            ]
+        }
+    },
+    {
+        "id": 31893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002229.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002229.wav"
+            ]
+        }
+    },
+    {
+        "id": 31897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001977.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001977.wav"
+            ]
+        }
+    },
+    {
+        "id": 31900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007251.wav"
+            ]
+        }
+    },
+    {
+        "id": 31901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003882.wav"
+            ]
+        }
+    },
+    {
+        "id": 31918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006382.wav"
+            ]
+        }
+    },
+    {
+        "id": 31919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005158.wav"
+            ]
+        }
+    },
+    {
+        "id": 31923,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right more rhythmic than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001188.wav"
+            ]
+        }
+    },
+    {
+        "id": 31928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003476.wav"
+            ]
+        }
+    },
+    {
+        "id": 31934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005831.wav"
+            ]
+        }
+    },
+    {
+        "id": 31948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006431.wav"
+            ]
+        }
+    },
+    {
+        "id": 31950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001545.wav"
+            ]
+        }
+    },
+    {
+        "id": 31955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007286.wav"
+            ]
+        }
+    },
+    {
+        "id": 31957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005096.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005096.wav"
+            ]
+        }
+    },
+    {
+        "id": 31960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003264.wav"
+            ]
+        }
+    },
+    {
+        "id": 31969,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000062.wav"
+            ]
+        }
+    },
+    {
+        "id": 31971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003278.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003278.wav"
+            ]
+        }
+    },
+    {
+        "id": 31972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008719.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008719.wav"
+            ]
+        }
+    },
+    {
+        "id": 31975,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002002.wav"
+            ]
+        }
+    },
+    {
+        "id": 31981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003572.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003572.wav"
+            ]
+        }
+    },
+    {
+        "id": 31993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000683.wav"
+            ]
+        }
+    },
+    {
+        "id": 31995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005449.wav"
+            ]
+        }
+    },
+    {
+        "id": 32002,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007273.wav"
+            ]
+        }
+    },
+    {
+        "id": 32003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006210.wav"
+            ]
+        }
+    },
+    {
+        "id": 32017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003318.wav"
+            ]
+        }
+    },
+    {
+        "id": 32018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007127.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007127.wav"
+            ]
+        }
+    },
+    {
+        "id": 32021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003627.wav"
+            ]
+        }
+    },
+    {
+        "id": 32029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006976.wav"
+            ]
+        }
+    },
+    {
+        "id": 32040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008623.wav"
+            ]
+        }
+    },
+    {
+        "id": 32043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002728.wav"
+            ]
+        }
+    },
+    {
+        "id": 32044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004156.wav"
+            ]
+        }
+    },
+    {
+        "id": 32046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000371.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000371.wav"
+            ]
+        }
+    },
+    {
+        "id": 32053,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005635.wav"
+            ]
+        }
+    },
+    {
+        "id": 32062,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004781.wav"
+            ]
+        }
+    },
+    {
+        "id": 32063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004079.wav"
+            ]
+        }
+    },
+    {
+        "id": 32075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006962.wav"
+            ]
+        }
+    },
+    {
+        "id": 32083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006513.wav"
+            ]
+        }
+    },
+    {
+        "id": 32086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005347.wav"
+            ]
+        }
+    },
+    {
+        "id": 32118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007886.wav"
+            ]
+        }
+    },
+    {
+        "id": 32120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005301.wav"
+            ]
+        }
+    },
+    {
+        "id": 32122,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008465.wav"
+            ]
+        }
+    },
+    {
+        "id": 32145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007466.wav"
+            ]
+        }
+    },
+    {
+        "id": 32146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000302.wav"
+            ]
+        }
+    },
+    {
+        "id": 32147,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006345.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006345.wav"
+            ]
+        }
+    },
+    {
+        "id": 32149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003065.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003065.wav"
+            ]
+        }
+    },
+    {
+        "id": 32155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000444.wav"
+            ]
+        }
+    },
+    {
+        "id": 32157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000404.wav"
+            ]
+        }
+    },
+    {
+        "id": 32160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005092.wav"
+            ]
+        }
+    },
+    {
+        "id": 32172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000930.wav"
+            ]
+        }
+    },
+    {
+        "id": 32180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006440.wav"
+            ]
+        }
+    },
+    {
+        "id": 32181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006455.wav"
+            ]
+        }
+    },
+    {
+        "id": 32183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004649.wav"
+            ]
+        }
+    },
+    {
+        "id": 32185,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005390.wav"
+            ]
+        }
+    },
+    {
+        "id": 32193,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003373.wav"
+            ]
+        }
+    },
+    {
+        "id": 32194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005599.wav"
+            ]
+        }
+    },
+    {
+        "id": 32214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002181.wav"
+            ]
+        }
+    },
+    {
+        "id": 32215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001667.wav"
+            ]
+        }
+    },
+    {
+        "id": 32221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002198.wav"
+            ]
+        }
+    },
+    {
+        "id": 32223,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003400.wav"
+            ]
+        }
+    },
+    {
+        "id": 32224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006725.wav"
+            ]
+        }
+    },
+    {
+        "id": 32234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000088.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000088.wav"
+            ]
+        }
+    },
+    {
+        "id": 32242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004748.wav"
+            ]
+        }
+    },
+    {
+        "id": 32243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005690.wav"
+            ]
+        }
+    },
+    {
+        "id": 32245,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002079.wav"
+            ]
+        }
+    },
+    {
+        "id": 32255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003093.wav"
+            ]
+        }
+    },
+    {
+        "id": 32256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003108.wav"
+            ]
+        }
+    },
+    {
+        "id": 32258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003185.wav"
+            ]
+        }
+    },
+    {
+        "id": 32263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002597.wav"
+            ]
+        }
+    },
+    {
+        "id": 32271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001225.wav"
+            ]
+        }
+    },
+    {
+        "id": 32280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004046.wav"
+            ]
+        }
+    },
+    {
+        "id": 32281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007519.wav"
+            ]
+        }
+    },
+    {
+        "id": 32284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008277.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008277.wav"
+            ]
+        }
+    },
+    {
+        "id": 32286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000039.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000039.wav"
+            ]
+        }
+    },
+    {
+        "id": 32289,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002546.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002546.wav"
+            ]
+        }
+    },
+    {
+        "id": 32291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006567.wav"
+            ]
+        }
+    },
+    {
+        "id": 32300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008418.wav"
+            ]
+        }
+    },
+    {
+        "id": 32313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005511.wav"
+            ]
+        }
+    },
+    {
+        "id": 32329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005455.wav"
+            ]
+        }
+    },
+    {
+        "id": 32340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007155.wav"
+            ]
+        }
+    },
+    {
+        "id": 32344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002310.wav"
+            ]
+        }
+    },
+    {
+        "id": 32346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000582.wav"
+            ]
+        }
+    },
+    {
+        "id": 32353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005715.wav"
+            ]
+        }
+    },
+    {
+        "id": 32360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004933.wav"
+            ]
+        }
+    },
+    {
+        "id": 32362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the banjo on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003613.wav"
+            ]
+        }
+    },
+    {
+        "id": 32363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000024.wav"
+            ]
+        }
+    },
+    {
+        "id": 32373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001429.wav"
+            ]
+        }
+    },
+    {
+        "id": 32375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005741.wav"
+            ]
+        }
+    },
+    {
+        "id": 32376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008835.wav"
+            ]
+        }
+    },
+    {
+        "id": 32378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005899.wav"
+            ]
+        }
+    },
+    {
+        "id": 32389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006162.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006162.wav"
+            ]
+        }
+    },
+    {
+        "id": 32393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003242.wav"
+            ]
+        }
+    },
+    {
+        "id": 32398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004603.wav"
+            ]
+        }
+    },
+    {
+        "id": 32402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003750.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003750.wav"
+            ]
+        }
+    },
+    {
+        "id": 32406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002055.wav"
+            ]
+        }
+    },
+    {
+        "id": 32409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007703.wav"
+            ]
+        }
+    },
+    {
+        "id": 32418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000264.wav"
+            ]
+        }
+    },
+    {
+        "id": 32428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007647.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007647.wav"
+            ]
+        }
+    },
+    {
+        "id": 32430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005961.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005961.wav"
+            ]
+        }
+    },
+    {
+        "id": 32435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000672.wav"
+            ]
+        }
+    },
+    {
+        "id": 32436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005221.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005221.wav"
+            ]
+        }
+    },
+    {
+        "id": 32438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005922.wav"
+            ]
+        }
+    },
+    {
+        "id": 32443,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001739.wav"
+            ]
+        }
+    },
+    {
+        "id": 32446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003219.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003219.wav"
+            ]
+        }
+    },
+    {
+        "id": 32448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000499.wav"
+            ]
+        }
+    },
+    {
+        "id": 32469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007846.wav"
+            ]
+        }
+    },
+    {
+        "id": 32479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006492.wav"
+            ]
+        }
+    },
+    {
+        "id": 32480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006944.wav"
+            ]
+        }
+    },
+    {
+        "id": 32485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001603.wav"
+            ]
+        }
+    },
+    {
+        "id": 32499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008072.wav"
+            ]
+        }
+    },
+    {
+        "id": 32506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002921.wav"
+            ]
+        }
+    },
+    {
+        "id": 32509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001234.wav"
+            ]
+        }
+    },
+    {
+        "id": 32510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005825.wav"
+            ]
+        }
+    },
+    {
+        "id": 32513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006631.wav"
+            ]
+        }
+    },
+    {
+        "id": 32516,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003822.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003822.wav"
+            ]
+        }
+    },
+    {
+        "id": 32519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003245.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003245.wav"
+            ]
+        }
+    },
+    {
+        "id": 32521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006168.wav"
+            ]
+        }
+    },
+    {
+        "id": 32529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005825.wav"
+            ]
+        }
+    },
+    {
+        "id": 32540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006271.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006271.wav"
+            ]
+        }
+    },
+    {
+        "id": 32542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000957.wav"
+            ]
+        }
+    },
+    {
+        "id": 32556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007735.wav"
+            ]
+        }
+    },
+    {
+        "id": 32557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001057.wav"
+            ]
+        }
+    },
+    {
+        "id": 32558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005958.wav"
+            ]
+        }
+    },
+    {
+        "id": 32564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003639.wav"
+            ]
+        }
+    },
+    {
+        "id": 32582,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008061.wav"
+            ]
+        }
+    },
+    {
+        "id": 32587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005708.wav"
+            ]
+        }
+    },
+    {
+        "id": 32592,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000363.wav"
+            ]
+        }
+    },
+    {
+        "id": 32593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003522.wav"
+            ]
+        }
+    },
+    {
+        "id": 32594,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000474.wav"
+            ]
+        }
+    },
+    {
+        "id": 32595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006564.wav"
+            ]
+        }
+    },
+    {
+        "id": 32597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008890.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008890.wav"
+            ]
+        }
+    },
+    {
+        "id": 32601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002449.wav"
+            ]
+        }
+    },
+    {
+        "id": 32602,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004199.wav"
+            ]
+        }
+    },
+    {
+        "id": 32622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006248.wav"
+            ]
+        }
+    },
+    {
+        "id": 32625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002154.wav"
+            ]
+        }
+    },
+    {
+        "id": 32634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007385.wav"
+            ]
+        }
+    },
+    {
+        "id": 32639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007543.wav"
+            ]
+        }
+    },
+    {
+        "id": 32640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001649.wav"
+            ]
+        }
+    },
+    {
+        "id": 32642,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002673.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002673.wav"
+            ]
+        }
+    },
+    {
+        "id": 32644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006010.wav"
+            ]
+        }
+    },
+    {
+        "id": 32645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005232.wav"
+            ]
+        }
+    },
+    {
+        "id": 32652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003176.wav"
+            ]
+        }
+    },
+    {
+        "id": 32658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008159.wav"
+            ]
+        }
+    },
+    {
+        "id": 32667,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005075.wav"
+            ]
+        }
+    },
+    {
+        "id": 32674,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002223.wav"
+            ]
+        }
+    },
+    {
+        "id": 32676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000122.wav"
+            ]
+        }
+    },
+    {
+        "id": 32677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003324.wav"
+            ]
+        }
+    },
+    {
+        "id": 32683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002784.wav"
+            ]
+        }
+    },
+    {
+        "id": 32689,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008927.wav"
+            ]
+        }
+    },
+    {
+        "id": 32699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008553.wav"
+            ]
+        }
+    },
+    {
+        "id": 32704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006966.wav"
+            ]
+        }
+    },
+    {
+        "id": 32710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006269.wav"
+            ]
+        }
+    },
+    {
+        "id": 32711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008407.wav"
+            ]
+        }
+    },
+    {
+        "id": 32713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004488.wav"
+            ]
+        }
+    },
+    {
+        "id": 32716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007399.wav"
+            ]
+        }
+    },
+    {
+        "id": 32718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000096.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000096.wav"
+            ]
+        }
+    },
+    {
+        "id": 32719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003209.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003209.wav"
+            ]
+        }
+    },
+    {
+        "id": 32724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006046.wav"
+            ]
+        }
+    },
+    {
+        "id": 32730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007422.wav"
+            ]
+        }
+    },
+    {
+        "id": 32734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006826.wav"
+            ]
+        }
+    },
+    {
+        "id": 32738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007134.wav"
+            ]
+        }
+    },
+    {
+        "id": 32744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002964.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002964.wav"
+            ]
+        }
+    },
+    {
+        "id": 32751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005784.wav"
+            ]
+        }
+    },
+    {
+        "id": 32754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004966.wav"
+            ]
+        }
+    },
+    {
+        "id": 32760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006225.wav"
+            ]
+        }
+    },
+    {
+        "id": 32764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002500.wav"
+            ]
+        }
+    },
+    {
+        "id": 32765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005255.wav"
+            ]
+        }
+    },
+    {
+        "id": 32768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007032.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007032.wav"
+            ]
+        }
+    },
+    {
+        "id": 32769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004154.wav"
+            ]
+        }
+    },
+    {
+        "id": 32777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004818.wav"
+            ]
+        }
+    },
+    {
+        "id": 32778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003034.wav"
+            ]
+        }
+    },
+    {
+        "id": 32789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008208.wav"
+            ]
+        }
+    },
+    {
+        "id": 32807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006904.wav"
+            ]
+        }
+    },
+    {
+        "id": 32812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008582.wav"
+            ]
+        }
+    },
+    {
+        "id": 32815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007509.wav"
+            ]
+        }
+    },
+    {
+        "id": 32824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001127.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001127.wav"
+            ]
+        }
+    },
+    {
+        "id": 32831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006489.wav"
+            ]
+        }
+    },
+    {
+        "id": 32834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008826.wav"
+            ]
+        }
+    },
+    {
+        "id": 32843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001574.wav"
+            ]
+        }
+    },
+    {
+        "id": 32844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000888.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000888.wav"
+            ]
+        }
+    },
+    {
+        "id": 32845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006417.wav"
+            ]
+        }
+    },
+    {
+        "id": 32847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003903.wav"
+            ]
+        }
+    },
+    {
+        "id": 32848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001945.wav"
+            ]
+        }
+    },
+    {
+        "id": 32851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005188.wav"
+            ]
+        }
+    },
+    {
+        "id": 32854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005070.wav"
+            ]
+        }
+    },
+    {
+        "id": 32855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003159.wav"
+            ]
+        }
+    },
+    {
+        "id": 32883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001307.wav"
+            ]
+        }
+    },
+    {
+        "id": 32885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002774.wav"
+            ]
+        }
+    },
+    {
+        "id": 32890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000453.wav"
+            ]
+        }
+    },
+    {
+        "id": 32891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002560.wav"
+            ]
+        }
+    },
+    {
+        "id": 32897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001771.wav"
+            ]
+        }
+    },
+    {
+        "id": 32899,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005854.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005854.wav"
+            ]
+        }
+    },
+    {
+        "id": 32901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001598.wav"
+            ]
+        }
+    },
+    {
+        "id": 32915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000860.wav"
+            ]
+        }
+    },
+    {
+        "id": 32936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006087.wav"
+            ]
+        }
+    },
+    {
+        "id": 32938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000180.wav"
+            ]
+        }
+    },
+    {
+        "id": 32950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005138.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005138.wav"
+            ]
+        }
+    },
+    {
+        "id": 32953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008215.wav"
+            ]
+        }
+    },
+    {
+        "id": 32954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004943.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004943.wav"
+            ]
+        }
+    },
+    {
+        "id": 32955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008851.wav"
+            ]
+        }
+    },
+    {
+        "id": 32958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005533.wav"
+            ]
+        }
+    },
+    {
+        "id": 32960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003748.wav"
+            ]
+        }
+    },
+    {
+        "id": 32972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002511.wav"
+            ]
+        }
+    },
+    {
+        "id": 32976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005179.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005179.wav"
+            ]
+        }
+    },
+    {
+        "id": 32977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008349.wav"
+            ]
+        }
+    },
+    {
+        "id": 32984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000161.wav"
+            ]
+        }
+    },
+    {
+        "id": 32985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005126.wav"
+            ]
+        }
+    },
+    {
+        "id": 33006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001712.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001712.wav"
+            ]
+        }
+    },
+    {
+        "id": 33008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005605.wav"
+            ]
+        }
+    },
+    {
+        "id": 33020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008607.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008607.wav"
+            ]
+        }
+    },
+    {
+        "id": 33026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006564.wav"
+            ]
+        }
+    },
+    {
+        "id": 33030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left louder than the pipa on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003264.wav"
+            ]
+        }
+    },
+    {
+        "id": 33031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004341.wav"
+            ]
+        }
+    },
+    {
+        "id": 33041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007376.wav"
+            ]
+        }
+    },
+    {
+        "id": 33061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008748.wav"
+            ]
+        }
+    },
+    {
+        "id": 33069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006931.wav"
+            ]
+        }
+    },
+    {
+        "id": 33075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003141.wav"
+            ]
+        }
+    },
+    {
+        "id": 33079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004520.wav"
+            ]
+        }
+    },
+    {
+        "id": 33082,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005063.wav"
+            ]
+        }
+    },
+    {
+        "id": 33083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000420.wav"
+            ]
+        }
+    },
+    {
+        "id": 33087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005796.wav"
+            ]
+        }
+    },
+    {
+        "id": 33088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006308.wav"
+            ]
+        }
+    },
+    {
+        "id": 33090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002129.wav"
+            ]
+        }
+    },
+    {
+        "id": 33094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004419.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004419.wav"
+            ]
+        }
+    },
+    {
+        "id": 33096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007091.wav"
+            ]
+        }
+    },
+    {
+        "id": 33099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004111.wav"
+            ]
+        }
+    },
+    {
+        "id": 33107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008413.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008413.wav"
+            ]
+        }
+    },
+    {
+        "id": 33121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007598.wav"
+            ]
+        }
+    },
+    {
+        "id": 33134,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004265.wav"
+            ]
+        }
+    },
+    {
+        "id": 33136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005480.wav"
+            ]
+        }
+    },
+    {
+        "id": 33137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003391.wav"
+            ]
+        }
+    },
+    {
+        "id": 33139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000813.wav"
+            ]
+        }
+    },
+    {
+        "id": 33143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006598.wav"
+            ]
+        }
+    },
+    {
+        "id": 33145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000969.wav"
+            ]
+        }
+    },
+    {
+        "id": 33148,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000497.wav"
+            ]
+        }
+    },
+    {
+        "id": 33156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005946.wav"
+            ]
+        }
+    },
+    {
+        "id": 33157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005134.wav"
+            ]
+        }
+    },
+    {
+        "id": 33159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007865.wav"
+            ]
+        }
+    },
+    {
+        "id": 33160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003798.wav"
+            ]
+        }
+    },
+    {
+        "id": 33162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005769.wav"
+            ]
+        }
+    },
+    {
+        "id": 33165,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000550.wav"
+            ]
+        }
+    },
+    {
+        "id": 33167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008893.wav"
+            ]
+        }
+    },
+    {
+        "id": 33175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008215.wav"
+            ]
+        }
+    },
+    {
+        "id": 33179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002562.wav"
+            ]
+        }
+    },
+    {
+        "id": 33181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004809.wav"
+            ]
+        }
+    },
+    {
+        "id": 33182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003081.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003081.wav"
+            ]
+        }
+    },
+    {
+        "id": 33183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000372.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000372.wav"
+            ]
+        }
+    },
+    {
+        "id": 33190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005708.wav"
+            ]
+        }
+    },
+    {
+        "id": 33196,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003294.wav"
+            ]
+        }
+    },
+    {
+        "id": 33198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006047.wav"
+            ]
+        }
+    },
+    {
+        "id": 33199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000668.wav"
+            ]
+        }
+    },
+    {
+        "id": 33210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001012.wav"
+            ]
+        }
+    },
+    {
+        "id": 33222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004105.wav"
+            ]
+        }
+    },
+    {
+        "id": 33227,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005434.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005434.wav"
+            ]
+        }
+    },
+    {
+        "id": 33230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003199.wav"
+            ]
+        }
+    },
+    {
+        "id": 33239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001670.wav"
+            ]
+        }
+    },
+    {
+        "id": 33240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004258.wav"
+            ]
+        }
+    },
+    {
+        "id": 33241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000860.wav"
+            ]
+        }
+    },
+    {
+        "id": 33246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006543.wav"
+            ]
+        }
+    },
+    {
+        "id": 33253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007588.wav"
+            ]
+        }
+    },
+    {
+        "id": 33270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003882.wav"
+            ]
+        }
+    },
+    {
+        "id": 33277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004758.wav"
+            ]
+        }
+    },
+    {
+        "id": 33280,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000788.wav"
+            ]
+        }
+    },
+    {
+        "id": 33281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000282.wav"
+            ]
+        }
+    },
+    {
+        "id": 33287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006971.wav"
+            ]
+        }
+    },
+    {
+        "id": 33298,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006669.wav"
+            ]
+        }
+    },
+    {
+        "id": 33310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007799.wav"
+            ]
+        }
+    },
+    {
+        "id": 33314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002511.wav"
+            ]
+        }
+    },
+    {
+        "id": 33319,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003775.wav"
+            ]
+        }
+    },
+    {
+        "id": 33321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000355.wav"
+            ]
+        }
+    },
+    {
+        "id": 33326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001168.wav"
+            ]
+        }
+    },
+    {
+        "id": 33327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003135.wav"
+            ]
+        }
+    },
+    {
+        "id": 33328,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002738.wav"
+            ]
+        }
+    },
+    {
+        "id": 33334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008281.wav"
+            ]
+        }
+    },
+    {
+        "id": 33335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007325.wav"
+            ]
+        }
+    },
+    {
+        "id": 33346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006078.wav"
+            ]
+        }
+    },
+    {
+        "id": 33348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005916.wav"
+            ]
+        }
+    },
+    {
+        "id": 33349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008601.wav"
+            ]
+        }
+    },
+    {
+        "id": 33351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005202.wav"
+            ]
+        }
+    },
+    {
+        "id": 33352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002617.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002617.wav"
+            ]
+        }
+    },
+    {
+        "id": 33355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000180.wav"
+            ]
+        }
+    },
+    {
+        "id": 33358,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004701.wav"
+            ]
+        }
+    },
+    {
+        "id": 33366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the leftest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002851.wav"
+            ]
+        }
+    },
+    {
+        "id": 33367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008579.wav"
+            ]
+        }
+    },
+    {
+        "id": 33369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007925.wav"
+            ]
+        }
+    },
+    {
+        "id": 33378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007613.wav"
+            ]
+        }
+    },
+    {
+        "id": 33382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003216.wav"
+            ]
+        }
+    },
+    {
+        "id": 33383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002639.wav"
+            ]
+        }
+    },
+    {
+        "id": 33387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003564.wav"
+            ]
+        }
+    },
+    {
+        "id": 33389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007405.wav"
+            ]
+        }
+    },
+    {
+        "id": 33401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005473.wav"
+            ]
+        }
+    },
+    {
+        "id": 33404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007504.wav"
+            ]
+        }
+    },
+    {
+        "id": 33416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002778.wav"
+            ]
+        }
+    },
+    {
+        "id": 33417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005079.wav"
+            ]
+        }
+    },
+    {
+        "id": 33420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005871.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005871.wav"
+            ]
+        }
+    },
+    {
+        "id": 33424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008365.wav"
+            ]
+        }
+    },
+    {
+        "id": 33429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001485.wav"
+            ]
+        }
+    },
+    {
+        "id": 33437,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003502.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003502.wav"
+            ]
+        }
+    },
+    {
+        "id": 33439,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001136.wav"
+            ]
+        }
+    },
+    {
+        "id": 33445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004714.wav"
+            ]
+        }
+    },
+    {
+        "id": 33446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001849.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001849.wav"
+            ]
+        }
+    },
+    {
+        "id": 33448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004417.wav"
+            ]
+        }
+    },
+    {
+        "id": 33452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005568.wav"
+            ]
+        }
+    },
+    {
+        "id": 33481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000653.wav"
+            ]
+        }
+    },
+    {
+        "id": 33483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002568.wav"
+            ]
+        }
+    },
+    {
+        "id": 33490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004142.wav"
+            ]
+        }
+    },
+    {
+        "id": 33498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000123.wav"
+            ]
+        }
+    },
+    {
+        "id": 33500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007108.wav"
+            ]
+        }
+    },
+    {
+        "id": 33511,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003059.wav"
+            ]
+        }
+    },
+    {
+        "id": 33514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008552.wav"
+            ]
+        }
+    },
+    {
+        "id": 33517,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006115.wav"
+            ]
+        }
+    },
+    {
+        "id": 33521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000693.wav"
+            ]
+        }
+    },
+    {
+        "id": 33524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006867.wav"
+            ]
+        }
+    },
+    {
+        "id": 33531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007469.wav"
+            ]
+        }
+    },
+    {
+        "id": 33534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006458.wav"
+            ]
+        }
+    },
+    {
+        "id": 33542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005899.wav"
+            ]
+        }
+    },
+    {
+        "id": 33550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003505.wav"
+            ]
+        }
+    },
+    {
+        "id": 33558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004623.wav"
+            ]
+        }
+    },
+    {
+        "id": 33561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of instrument is the rightest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004210.wav"
+            ]
+        }
+    },
+    {
+        "id": 33564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002334.wav"
+            ]
+        }
+    },
+    {
+        "id": 33566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007748.wav"
+            ]
+        }
+    },
+    {
+        "id": 33573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002108.wav"
+            ]
+        }
+    },
+    {
+        "id": 33575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001287.wav"
+            ]
+        }
+    },
+    {
+        "id": 33580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004034.wav"
+            ]
+        }
+    },
+    {
+        "id": 33588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006161.wav"
+            ]
+        }
+    },
+    {
+        "id": 33591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008250.wav"
+            ]
+        }
+    },
+    {
+        "id": 33592,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007469.wav"
+            ]
+        }
+    },
+    {
+        "id": 33605,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001354.wav"
+            ]
+        }
+    },
+    {
+        "id": 33607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000383.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000383.wav"
+            ]
+        }
+    },
+    {
+        "id": 33609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006143.wav"
+            ]
+        }
+    },
+    {
+        "id": 33613,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002143.wav"
+            ]
+        }
+    },
+    {
+        "id": 33616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003456.wav"
+            ]
+        }
+    },
+    {
+        "id": 33620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007774.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007774.wav"
+            ]
+        }
+    },
+    {
+        "id": 33621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002708.wav"
+            ]
+        }
+    },
+    {
+        "id": 33624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004117.wav"
+            ]
+        }
+    },
+    {
+        "id": 33627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008257.wav"
+            ]
+        }
+    },
+    {
+        "id": 33630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000354.wav"
+            ]
+        }
+    },
+    {
+        "id": 33631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003895.wav"
+            ]
+        }
+    },
+    {
+        "id": 33645,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005211.wav"
+            ]
+        }
+    },
+    {
+        "id": 33651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008022.wav"
+            ]
+        }
+    },
+    {
+        "id": 33652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008852.wav"
+            ]
+        }
+    },
+    {
+        "id": 33670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008553.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008553.wav"
+            ]
+        }
+    },
+    {
+        "id": 33671,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002547.wav"
+            ]
+        }
+    },
+    {
+        "id": 33679,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004277.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004277.wav"
+            ]
+        }
+    },
+    {
+        "id": 33684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000487.wav"
+            ]
+        }
+    },
+    {
+        "id": 33686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005765.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005765.wav"
+            ]
+        }
+    },
+    {
+        "id": 33687,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008754.wav"
+            ]
+        }
+    },
+    {
+        "id": 33689,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001870.wav"
+            ]
+        }
+    },
+    {
+        "id": 33700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007800.wav"
+            ]
+        }
+    },
+    {
+        "id": 33702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005238.wav"
+            ]
+        }
+    },
+    {
+        "id": 33708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004296.wav"
+            ]
+        }
+    },
+    {
+        "id": 33718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007171.wav"
+            ]
+        }
+    },
+    {
+        "id": 33719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003668.wav"
+            ]
+        }
+    },
+    {
+        "id": 33721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008411.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008411.wav"
+            ]
+        }
+    },
+    {
+        "id": 33725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007956.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007956.wav"
+            ]
+        }
+    },
+    {
+        "id": 33747,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001994.wav"
+            ]
+        }
+    },
+    {
+        "id": 33748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006446.wav"
+            ]
+        }
+    },
+    {
+        "id": 33750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008039.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008039.wav"
+            ]
+        }
+    },
+    {
+        "id": 33751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006480.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006480.wav"
+            ]
+        }
+    },
+    {
+        "id": 33754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005912.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005912.wav"
+            ]
+        }
+    },
+    {
+        "id": 33756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003222.wav"
+            ]
+        }
+    },
+    {
+        "id": 33758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000508.wav"
+            ]
+        }
+    },
+    {
+        "id": 33762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000711.wav"
+            ]
+        }
+    },
+    {
+        "id": 33763,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004569.wav"
+            ]
+        }
+    },
+    {
+        "id": 33769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004154.wav"
+            ]
+        }
+    },
+    {
+        "id": 33781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000268.wav"
+            ]
+        }
+    },
+    {
+        "id": 33791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001543.wav"
+            ]
+        }
+    },
+    {
+        "id": 33792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004009.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004009.wav"
+            ]
+        }
+    },
+    {
+        "id": 33796,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007925.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007925.wav"
+            ]
+        }
+    },
+    {
+        "id": 33808,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008506.wav"
+            ]
+        }
+    },
+    {
+        "id": 33809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003654.wav"
+            ]
+        }
+    },
+    {
+        "id": 33814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006721.wav"
+            ]
+        }
+    },
+    {
+        "id": 33822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007524.wav"
+            ]
+        }
+    },
+    {
+        "id": 33824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008704.wav"
+            ]
+        }
+    },
+    {
+        "id": 33832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000656.wav"
+            ]
+        }
+    },
+    {
+        "id": 33834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004746.wav"
+            ]
+        }
+    },
+    {
+        "id": 33836,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006889.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006889.wav"
+            ]
+        }
+    },
+    {
+        "id": 33848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002255.wav"
+            ]
+        }
+    },
+    {
+        "id": 33849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004828.wav"
+            ]
+        }
+    },
+    {
+        "id": 33850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003229.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003229.wav"
+            ]
+        }
+    },
+    {
+        "id": 33858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004592.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004592.wav"
+            ]
+        }
+    },
+    {
+        "id": 33861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002343.wav"
+            ]
+        }
+    },
+    {
+        "id": 33868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002414.wav"
+            ]
+        }
+    },
+    {
+        "id": 33870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002269.wav"
+            ]
+        }
+    },
+    {
+        "id": 33871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007058.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007058.wav"
+            ]
+        }
+    },
+    {
+        "id": 33875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000401.wav"
+            ]
+        }
+    },
+    {
+        "id": 33876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007310.wav"
+            ]
+        }
+    },
+    {
+        "id": 33881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005681.wav"
+            ]
+        }
+    },
+    {
+        "id": 33883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000295.wav"
+            ]
+        }
+    },
+    {
+        "id": 33886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003181.wav"
+            ]
+        }
+    },
+    {
+        "id": 33889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005342.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005342.wav"
+            ]
+        }
+    },
+    {
+        "id": 33902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002771.wav"
+            ]
+        }
+    },
+    {
+        "id": 33903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001028.wav"
+            ]
+        }
+    },
+    {
+        "id": 33907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006820.wav"
+            ]
+        }
+    },
+    {
+        "id": 33914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002063.wav"
+            ]
+        }
+    },
+    {
+        "id": 33921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004791.wav"
+            ]
+        }
+    },
+    {
+        "id": 33935,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001797.wav"
+            ]
+        }
+    },
+    {
+        "id": 33948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001200.wav"
+            ]
+        }
+    },
+    {
+        "id": 33956,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006994.wav"
+            ]
+        }
+    },
+    {
+        "id": 33985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008270.wav"
+            ]
+        }
+    },
+    {
+        "id": 33986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004350.wav"
+            ]
+        }
+    },
+    {
+        "id": 33990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002479.wav"
+            ]
+        }
+    },
+    {
+        "id": 33991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003010.wav"
+            ]
+        }
+    },
+    {
+        "id": 33993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001037.wav"
+            ]
+        }
+    },
+    {
+        "id": 33998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005987.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005987.wav"
+            ]
+        }
+    },
+    {
+        "id": 34022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005819.wav"
+            ]
+        }
+    },
+    {
+        "id": 34036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005477.wav"
+            ]
+        }
+    },
+    {
+        "id": 34040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004464.wav"
+            ]
+        }
+    },
+    {
+        "id": 34041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001606.wav"
+            ]
+        }
+    },
+    {
+        "id": 34043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007368.wav"
+            ]
+        }
+    },
+    {
+        "id": 34044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006910.wav"
+            ]
+        }
+    },
+    {
+        "id": 34047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008388.wav"
+            ]
+        }
+    },
+    {
+        "id": 34057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich acoustic_guitar makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003208.wav"
+            ]
+        }
+    },
+    {
+        "id": 34061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001924.wav"
+            ]
+        }
+    },
+    {
+        "id": 34064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000450.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000450.wav"
+            ]
+        }
+    },
+    {
+        "id": 34072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002564.wav"
+            ]
+        }
+    },
+    {
+        "id": 34073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003608.wav"
+            ]
+        }
+    },
+    {
+        "id": 34074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004265.wav"
+            ]
+        }
+    },
+    {
+        "id": 34075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003439.wav"
+            ]
+        }
+    },
+    {
+        "id": 34080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005788.wav"
+            ]
+        }
+    },
+    {
+        "id": 34088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003383.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003383.wav"
+            ]
+        }
+    },
+    {
+        "id": 34093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006587.wav"
+            ]
+        }
+    },
+    {
+        "id": 34097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000119.wav"
+            ]
+        }
+    },
+    {
+        "id": 34103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001155.wav"
+            ]
+        }
+    },
+    {
+        "id": 34104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007389.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007389.wav"
+            ]
+        }
+    },
+    {
+        "id": 34105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002403.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002403.wav"
+            ]
+        }
+    },
+    {
+        "id": 34112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003047.wav"
+            ]
+        }
+    },
+    {
+        "id": 34113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007129.wav"
+            ]
+        }
+    },
+    {
+        "id": 34116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000630.wav"
+            ]
+        }
+    },
+    {
+        "id": 34120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004384.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004384.wav"
+            ]
+        }
+    },
+    {
+        "id": 34127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005947.wav"
+            ]
+        }
+    },
+    {
+        "id": 34137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006736.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006736.wav"
+            ]
+        }
+    },
+    {
+        "id": 34138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003786.wav"
+            ]
+        }
+    },
+    {
+        "id": 34143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001241.wav"
+            ]
+        }
+    },
+    {
+        "id": 34145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002934.wav"
+            ]
+        }
+    },
+    {
+        "id": 34150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007416.wav"
+            ]
+        }
+    },
+    {
+        "id": 34154,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002269.wav"
+            ]
+        }
+    },
+    {
+        "id": 34158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002933.wav"
+            ]
+        }
+    },
+    {
+        "id": 34159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004148.wav"
+            ]
+        }
+    },
+    {
+        "id": 34164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000401.wav"
+            ]
+        }
+    },
+    {
+        "id": 34166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001166.wav"
+            ]
+        }
+    },
+    {
+        "id": 34167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004051.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004051.wav"
+            ]
+        }
+    },
+    {
+        "id": 34169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005171.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005171.wav"
+            ]
+        }
+    },
+    {
+        "id": 34175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007045.wav"
+            ]
+        }
+    },
+    {
+        "id": 34183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007319.wav"
+            ]
+        }
+    },
+    {
+        "id": 34187,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003196.wav"
+            ]
+        }
+    },
+    {
+        "id": 34188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000098.wav"
+            ]
+        }
+    },
+    {
+        "id": 34194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002852.wav"
+            ]
+        }
+    },
+    {
+        "id": 34196,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002989.wav"
+            ]
+        }
+    },
+    {
+        "id": 34199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006503.wav"
+            ]
+        }
+    },
+    {
+        "id": 34203,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006299.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006299.wav"
+            ]
+        }
+    },
+    {
+        "id": 34204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003152.wav"
+            ]
+        }
+    },
+    {
+        "id": 34218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008142.wav"
+            ]
+        }
+    },
+    {
+        "id": 34232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005515.wav"
+            ]
+        }
+    },
+    {
+        "id": 34237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002373.wav"
+            ]
+        }
+    },
+    {
+        "id": 34240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000639.wav"
+            ]
+        }
+    },
+    {
+        "id": 34242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008436.wav"
+            ]
+        }
+    },
+    {
+        "id": 34243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006063.wav"
+            ]
+        }
+    },
+    {
+        "id": 34255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007368.wav"
+            ]
+        }
+    },
+    {
+        "id": 34273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008714.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008714.wav"
+            ]
+        }
+    },
+    {
+        "id": 34274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000617.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000617.wav"
+            ]
+        }
+    },
+    {
+        "id": 34275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005843.wav"
+            ]
+        }
+    },
+    {
+        "id": 34278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004261.wav"
+            ]
+        }
+    },
+    {
+        "id": 34286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005037.wav"
+            ]
+        }
+    },
+    {
+        "id": 34310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007646.wav"
+            ]
+        }
+    },
+    {
+        "id": 34311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000323.wav"
+            ]
+        }
+    },
+    {
+        "id": 34312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005337.wav"
+            ]
+        }
+    },
+    {
+        "id": 34317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008167.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008167.wav"
+            ]
+        }
+    },
+    {
+        "id": 34318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006519.wav"
+            ]
+        }
+    },
+    {
+        "id": 34320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002340.wav"
+            ]
+        }
+    },
+    {
+        "id": 34323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007042.wav"
+            ]
+        }
+    },
+    {
+        "id": 34329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004057.wav"
+            ]
+        }
+    },
+    {
+        "id": 34332,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002451.wav"
+            ]
+        }
+    },
+    {
+        "id": 34333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002462.wav"
+            ]
+        }
+    },
+    {
+        "id": 34338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004753.wav"
+            ]
+        }
+    },
+    {
+        "id": 34341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002054.wav"
+            ]
+        }
+    },
+    {
+        "id": 34345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004909.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004909.wav"
+            ]
+        }
+    },
+    {
+        "id": 34349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007417.wav"
+            ]
+        }
+    },
+    {
+        "id": 34354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005293.wav"
+            ]
+        }
+    },
+    {
+        "id": 34356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007512.wav"
+            ]
+        }
+    },
+    {
+        "id": 34358,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003748.wav"
+            ]
+        }
+    },
+    {
+        "id": 34361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003414.wav"
+            ]
+        }
+    },
+    {
+        "id": 34369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002473.wav"
+            ]
+        }
+    },
+    {
+        "id": 34373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001295.wav"
+            ]
+        }
+    },
+    {
+        "id": 34377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003265.wav"
+            ]
+        }
+    },
+    {
+        "id": 34381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008084.wav"
+            ]
+        }
+    },
+    {
+        "id": 34384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005302.wav"
+            ]
+        }
+    },
+    {
+        "id": 34389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004335.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004335.wav"
+            ]
+        }
+    },
+    {
+        "id": 34390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003833.wav"
+            ]
+        }
+    },
+    {
+        "id": 34392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002328.wav"
+            ]
+        }
+    },
+    {
+        "id": 34411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001475.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001475.wav"
+            ]
+        }
+    },
+    {
+        "id": 34417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006547.wav"
+            ]
+        }
+    },
+    {
+        "id": 34422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006026.wav"
+            ]
+        }
+    },
+    {
+        "id": 34425,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003827.wav"
+            ]
+        }
+    },
+    {
+        "id": 34429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007895.wav"
+            ]
+        }
+    },
+    {
+        "id": 34438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000558.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000558.wav"
+            ]
+        }
+    },
+    {
+        "id": 34445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003005.wav"
+            ]
+        }
+    },
+    {
+        "id": 34457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002509.wav"
+            ]
+        }
+    },
+    {
+        "id": 34460,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006642.wav"
+            ]
+        }
+    },
+    {
+        "id": 34477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004715.wav"
+            ]
+        }
+    },
+    {
+        "id": 34486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006422.wav"
+            ]
+        }
+    },
+    {
+        "id": 34489,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006684.wav"
+            ]
+        }
+    },
+    {
+        "id": 34492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004799.wav"
+            ]
+        }
+    },
+    {
+        "id": 34496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003835.wav"
+            ]
+        }
+    },
+    {
+        "id": 34502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004288.wav"
+            ]
+        }
+    },
+    {
+        "id": 34503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002637.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002637.wav"
+            ]
+        }
+    },
+    {
+        "id": 34506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000063.wav"
+            ]
+        }
+    },
+    {
+        "id": 34521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008207.wav"
+            ]
+        }
+    },
+    {
+        "id": 34532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008828.wav"
+            ]
+        }
+    },
+    {
+        "id": 34541,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002607.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002607.wav"
+            ]
+        }
+    },
+    {
+        "id": 34542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005746.wav"
+            ]
+        }
+    },
+    {
+        "id": 34544,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007382.wav"
+            ]
+        }
+    },
+    {
+        "id": 34545,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001140.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001140.wav"
+            ]
+        }
+    },
+    {
+        "id": 34566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004166.wav"
+            ]
+        }
+    },
+    {
+        "id": 34569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006212.wav"
+            ]
+        }
+    },
+    {
+        "id": 34575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007892.wav"
+            ]
+        }
+    },
+    {
+        "id": 34577,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006494.wav"
+            ]
+        }
+    },
+    {
+        "id": 34580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003780.wav"
+            ]
+        }
+    },
+    {
+        "id": 34584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004188.wav"
+            ]
+        }
+    },
+    {
+        "id": 34585,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000621.wav"
+            ]
+        }
+    },
+    {
+        "id": 34586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006237.wav"
+            ]
+        }
+    },
+    {
+        "id": 34587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008462.wav"
+            ]
+        }
+    },
+    {
+        "id": 34600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007584.wav"
+            ]
+        }
+    },
+    {
+        "id": 34606,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008673.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008673.wav"
+            ]
+        }
+    },
+    {
+        "id": 34610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008329.wav"
+            ]
+        }
+    },
+    {
+        "id": 34615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006423.wav"
+            ]
+        }
+    },
+    {
+        "id": 34618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003725.wav"
+            ]
+        }
+    },
+    {
+        "id": 34636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008371.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008371.wav"
+            ]
+        }
+    },
+    {
+        "id": 34639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007726.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007726.wav"
+            ]
+        }
+    },
+    {
+        "id": 34652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007892.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007892.wav"
+            ]
+        }
+    },
+    {
+        "id": 34656,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004470.wav"
+            ]
+        }
+    },
+    {
+        "id": 34659,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001412.wav"
+            ]
+        }
+    },
+    {
+        "id": 34664,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006286.wav"
+            ]
+        }
+    },
+    {
+        "id": 34672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006824.wav"
+            ]
+        }
+    },
+    {
+        "id": 34677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005428.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005428.wav"
+            ]
+        }
+    },
+    {
+        "id": 34678,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002769.wav"
+            ]
+        }
+    },
+    {
+        "id": 34688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002699.wav"
+            ]
+        }
+    },
+    {
+        "id": 34695,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003627.wav"
+            ]
+        }
+    },
+    {
+        "id": 34699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the right more rhythmic than the guzheng on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001992.wav"
+            ]
+        }
+    },
+    {
+        "id": 34700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008279.wav"
+            ]
+        }
+    },
+    {
+        "id": 34701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005283.wav"
+            ]
+        }
+    },
+    {
+        "id": 34703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006537.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006537.wav"
+            ]
+        }
+    },
+    {
+        "id": 34708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000233.wav"
+            ]
+        }
+    },
+    {
+        "id": 34711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007865.wav"
+            ]
+        }
+    },
+    {
+        "id": 34712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006905.wav"
+            ]
+        }
+    },
+    {
+        "id": 34715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000731.wav"
+            ]
+        }
+    },
+    {
+        "id": 34716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002676.wav"
+            ]
+        }
+    },
+    {
+        "id": 34719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004829.wav"
+            ]
+        }
+    },
+    {
+        "id": 34721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004305.wav"
+            ]
+        }
+    },
+    {
+        "id": 34725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005536.wav"
+            ]
+        }
+    },
+    {
+        "id": 34729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004481.wav"
+            ]
+        }
+    },
+    {
+        "id": 34731,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008284.wav"
+            ]
+        }
+    },
+    {
+        "id": 34735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005118.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005118.wav"
+            ]
+        }
+    },
+    {
+        "id": 34736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006604.wav"
+            ]
+        }
+    },
+    {
+        "id": 34746,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003608.wav"
+            ]
+        }
+    },
+    {
+        "id": 34748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000830.wav"
+            ]
+        }
+    },
+    {
+        "id": 34750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003911.wav"
+            ]
+        }
+    },
+    {
+        "id": 34754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002793.wav"
+            ]
+        }
+    },
+    {
+        "id": 34756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002354.wav"
+            ]
+        }
+    },
+    {
+        "id": 34759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001992.wav"
+            ]
+        }
+    },
+    {
+        "id": 34767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008689.wav"
+            ]
+        }
+    },
+    {
+        "id": 34769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005583.wav"
+            ]
+        }
+    },
+    {
+        "id": 34781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002072.wav"
+            ]
+        }
+    },
+    {
+        "id": 34791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002934.wav"
+            ]
+        }
+    },
+    {
+        "id": 34793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008177.wav"
+            ]
+        }
+    },
+    {
+        "id": 34805,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004954.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004954.wav"
+            ]
+        }
+    },
+    {
+        "id": 34807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000952.wav"
+            ]
+        }
+    },
+    {
+        "id": 34812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002285.wav"
+            ]
+        }
+    },
+    {
+        "id": 34817,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003658.wav"
+            ]
+        }
+    },
+    {
+        "id": 34821,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008452.wav"
+            ]
+        }
+    },
+    {
+        "id": 34822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007886.wav"
+            ]
+        }
+    },
+    {
+        "id": 34828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006236.wav"
+            ]
+        }
+    },
+    {
+        "id": 34833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004204.wav"
+            ]
+        }
+    },
+    {
+        "id": 34836,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006758.wav"
+            ]
+        }
+    },
+    {
+        "id": 34841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005408.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005408.wav"
+            ]
+        }
+    },
+    {
+        "id": 34844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001251.wav"
+            ]
+        }
+    },
+    {
+        "id": 34853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003813.wav"
+            ]
+        }
+    },
+    {
+        "id": 34861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006396.wav"
+            ]
+        }
+    },
+    {
+        "id": 34866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005939.wav"
+            ]
+        }
+    },
+    {
+        "id": 34867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000634.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000634.wav"
+            ]
+        }
+    },
+    {
+        "id": 34876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005150.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005150.wav"
+            ]
+        }
+    },
+    {
+        "id": 34881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008632.wav"
+            ]
+        }
+    },
+    {
+        "id": 34891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006560.wav"
+            ]
+        }
+    },
+    {
+        "id": 34894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001456.wav"
+            ]
+        }
+    },
+    {
+        "id": 34895,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003817.wav"
+            ]
+        }
+    },
+    {
+        "id": 34907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000248.wav"
+            ]
+        }
+    },
+    {
+        "id": 34911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass on the left louder than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002223.wav"
+            ]
+        }
+    },
+    {
+        "id": 34920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008575.wav"
+            ]
+        }
+    },
+    {
+        "id": 34928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006753.wav"
+            ]
+        }
+    },
+    {
+        "id": 34936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004193.wav"
+            ]
+        }
+    },
+    {
+        "id": 34939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001292.wav"
+            ]
+        }
+    },
+    {
+        "id": 34940,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006738.wav"
+            ]
+        }
+    },
+    {
+        "id": 34943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008851.wav"
+            ]
+        }
+    },
+    {
+        "id": 34955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003102.wav"
+            ]
+        }
+    },
+    {
+        "id": 34958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003920.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003920.wav"
+            ]
+        }
+    },
+    {
+        "id": 34962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008471.wav"
+            ]
+        }
+    },
+    {
+        "id": 34966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002791.wav"
+            ]
+        }
+    },
+    {
+        "id": 34973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002536.wav"
+            ]
+        }
+    },
+    {
+        "id": 34978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008292.wav"
+            ]
+        }
+    },
+    {
+        "id": 34981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007695.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007695.wav"
+            ]
+        }
+    },
+    {
+        "id": 34982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002768.wav"
+            ]
+        }
+    },
+    {
+        "id": 34990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008870.wav"
+            ]
+        }
+    },
+    {
+        "id": 35000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008006.wav"
+            ]
+        }
+    },
+    {
+        "id": 35002,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006301.wav"
+            ]
+        }
+    },
+    {
+        "id": 35004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005783.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005783.wav"
+            ]
+        }
+    },
+    {
+        "id": 35005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003217.wav"
+            ]
+        }
+    },
+    {
+        "id": 35009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007711.wav"
+            ]
+        }
+    },
+    {
+        "id": 35010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005900.wav"
+            ]
+        }
+    },
+    {
+        "id": 35013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001584.wav"
+            ]
+        }
+    },
+    {
+        "id": 35014,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003457.wav"
+            ]
+        }
+    },
+    {
+        "id": 35017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005648.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005648.wav"
+            ]
+        }
+    },
+    {
+        "id": 35026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005881.wav"
+            ]
+        }
+    },
+    {
+        "id": 35028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006289.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006289.wav"
+            ]
+        }
+    },
+    {
+        "id": 35034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002749.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002749.wav"
+            ]
+        }
+    },
+    {
+        "id": 35039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006235.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006235.wav"
+            ]
+        }
+    },
+    {
+        "id": 35050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003586.wav"
+            ]
+        }
+    },
+    {
+        "id": 35051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007484.wav"
+            ]
+        }
+    },
+    {
+        "id": 35058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002549.wav"
+            ]
+        }
+    },
+    {
+        "id": 35060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002625.wav"
+            ]
+        }
+    },
+    {
+        "id": 35066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002784.wav"
+            ]
+        }
+    },
+    {
+        "id": 35078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005210.wav"
+            ]
+        }
+    },
+    {
+        "id": 35081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006944.wav"
+            ]
+        }
+    },
+    {
+        "id": 35090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000420.wav"
+            ]
+        }
+    },
+    {
+        "id": 35091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004513.wav"
+            ]
+        }
+    },
+    {
+        "id": 35093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005488.wav"
+            ]
+        }
+    },
+    {
+        "id": 35094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000756.wav"
+            ]
+        }
+    },
+    {
+        "id": 35097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000122.wav"
+            ]
+        }
+    },
+    {
+        "id": 35099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007968.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007968.wav"
+            ]
+        }
+    },
+    {
+        "id": 35100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002047.wav"
+            ]
+        }
+    },
+    {
+        "id": 35102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002882.wav"
+            ]
+        }
+    },
+    {
+        "id": 35103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005449.wav"
+            ]
+        }
+    },
+    {
+        "id": 35116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006291.wav"
+            ]
+        }
+    },
+    {
+        "id": 35121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008479.wav"
+            ]
+        }
+    },
+    {
+        "id": 35129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005005.wav"
+            ]
+        }
+    },
+    {
+        "id": 35131,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004788.wav"
+            ]
+        }
+    },
+    {
+        "id": 35137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000524.wav"
+            ]
+        }
+    },
+    {
+        "id": 35138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007272.wav"
+            ]
+        }
+    },
+    {
+        "id": 35146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005362.wav"
+            ]
+        }
+    },
+    {
+        "id": 35158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008761.wav"
+            ]
+        }
+    },
+    {
+        "id": 35163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001490.wav"
+            ]
+        }
+    },
+    {
+        "id": 35174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008770.wav"
+            ]
+        }
+    },
+    {
+        "id": 35176,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003741.wav"
+            ]
+        }
+    },
+    {
+        "id": 35181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002022.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002022.wav"
+            ]
+        }
+    },
+    {
+        "id": 35194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007796.wav"
+            ]
+        }
+    },
+    {
+        "id": 35216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005372.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005372.wav"
+            ]
+        }
+    },
+    {
+        "id": 35217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006610.wav"
+            ]
+        }
+    },
+    {
+        "id": 35219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002797.wav"
+            ]
+        }
+    },
+    {
+        "id": 35236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002579.wav"
+            ]
+        }
+    },
+    {
+        "id": 35237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006770.wav"
+            ]
+        }
+    },
+    {
+        "id": 35245,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008250.wav"
+            ]
+        }
+    },
+    {
+        "id": 35257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005762.wav"
+            ]
+        }
+    },
+    {
+        "id": 35266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich tuba makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002539.wav"
+            ]
+        }
+    },
+    {
+        "id": 35271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008092.wav"
+            ]
+        }
+    },
+    {
+        "id": 35273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007154.wav"
+            ]
+        }
+    },
+    {
+        "id": 35275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005580.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005580.wav"
+            ]
+        }
+    },
+    {
+        "id": 35281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000817.wav"
+            ]
+        }
+    },
+    {
+        "id": 35282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003185.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003185.wav"
+            ]
+        }
+    },
+    {
+        "id": 35283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006762.wav"
+            ]
+        }
+    },
+    {
+        "id": 35296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008829.wav"
+            ]
+        }
+    },
+    {
+        "id": 35301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003662.wav"
+            ]
+        }
+    },
+    {
+        "id": 35305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the saxophone on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006528.wav"
+            ]
+        }
+    },
+    {
+        "id": 35306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008070.wav"
+            ]
+        }
+    },
+    {
+        "id": 35307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004465.wav"
+            ]
+        }
+    },
+    {
+        "id": 35309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007112.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007112.wav"
+            ]
+        }
+    },
+    {
+        "id": 35312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000303.wav"
+            ]
+        }
+    },
+    {
+        "id": 35315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001965.wav"
+            ]
+        }
+    },
+    {
+        "id": 35322,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007696.wav"
+            ]
+        }
+    },
+    {
+        "id": 35323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006454.wav"
+            ]
+        }
+    },
+    {
+        "id": 35329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003539.wav"
+            ]
+        }
+    },
+    {
+        "id": 35330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000394.wav"
+            ]
+        }
+    },
+    {
+        "id": 35335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000233.wav"
+            ]
+        }
+    },
+    {
+        "id": 35339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008841.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008841.wav"
+            ]
+        }
+    },
+    {
+        "id": 35341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003123.wav"
+            ]
+        }
+    },
+    {
+        "id": 35346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007445.wav"
+            ]
+        }
+    },
+    {
+        "id": 35348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002470.wav"
+            ]
+        }
+    },
+    {
+        "id": 35351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003929.wav"
+            ]
+        }
+    },
+    {
+        "id": 35352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005410.wav"
+            ]
+        }
+    },
+    {
+        "id": 35361,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005684.wav"
+            ]
+        }
+    },
+    {
+        "id": 35363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007015.wav"
+            ]
+        }
+    },
+    {
+        "id": 35373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004652.wav"
+            ]
+        }
+    },
+    {
+        "id": 35384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003445.wav"
+            ]
+        }
+    },
+    {
+        "id": 35393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001014.wav"
+            ]
+        }
+    },
+    {
+        "id": 35396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004302.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004302.wav"
+            ]
+        }
+    },
+    {
+        "id": 35401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006393.wav"
+            ]
+        }
+    },
+    {
+        "id": 35405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007973.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007973.wav"
+            ]
+        }
+    },
+    {
+        "id": 35410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003833.wav"
+            ]
+        }
+    },
+    {
+        "id": 35414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000809.wav"
+            ]
+        }
+    },
+    {
+        "id": 35421,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007080.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007080.wav"
+            ]
+        }
+    },
+    {
+        "id": 35424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000102.wav"
+            ]
+        }
+    },
+    {
+        "id": 35441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002059.wav"
+            ]
+        }
+    },
+    {
+        "id": 35457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003460.wav"
+            ]
+        }
+    },
+    {
+        "id": 35459,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005337.wav"
+            ]
+        }
+    },
+    {
+        "id": 35461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003886.wav"
+            ]
+        }
+    },
+    {
+        "id": 35464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004205.wav"
+            ]
+        }
+    },
+    {
+        "id": 35465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004328.wav"
+            ]
+        }
+    },
+    {
+        "id": 35469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004182.wav"
+            ]
+        }
+    },
+    {
+        "id": 35471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004983.wav"
+            ]
+        }
+    },
+    {
+        "id": 35481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005805.wav"
+            ]
+        }
+    },
+    {
+        "id": 35483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007319.wav"
+            ]
+        }
+    },
+    {
+        "id": 35484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004363.wav"
+            ]
+        }
+    },
+    {
+        "id": 35487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005441.wav"
+            ]
+        }
+    },
+    {
+        "id": 35494,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007623.wav"
+            ]
+        }
+    },
+    {
+        "id": 35499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000829.wav"
+            ]
+        }
+    },
+    {
+        "id": 35510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002984.wav"
+            ]
+        }
+    },
+    {
+        "id": 35521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004976.wav"
+            ]
+        }
+    },
+    {
+        "id": 35529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004905.wav"
+            ]
+        }
+    },
+    {
+        "id": 35533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001624.wav"
+            ]
+        }
+    },
+    {
+        "id": 35540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008970.wav"
+            ]
+        }
+    },
+    {
+        "id": 35550,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005093.wav"
+            ]
+        }
+    },
+    {
+        "id": 35568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the right louder than the guzheng on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001797.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001797.wav"
+            ]
+        }
+    },
+    {
+        "id": 35569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007987.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007987.wav"
+            ]
+        }
+    },
+    {
+        "id": 35571,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005564.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005564.wav"
+            ]
+        }
+    },
+    {
+        "id": 35572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001522.wav"
+            ]
+        }
+    },
+    {
+        "id": 35579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002329.wav"
+            ]
+        }
+    },
+    {
+        "id": 35580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008563.wav"
+            ]
+        }
+    },
+    {
+        "id": 35587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000565.wav"
+            ]
+        }
+    },
+    {
+        "id": 35588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007043.wav"
+            ]
+        }
+    },
+    {
+        "id": 35589,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002544.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002544.wav"
+            ]
+        }
+    },
+    {
+        "id": 35592,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003770.wav"
+            ]
+        }
+    },
+    {
+        "id": 35600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008377.wav"
+            ]
+        }
+    },
+    {
+        "id": 35603,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008837.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008837.wav"
+            ]
+        }
+    },
+    {
+        "id": 35620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007469.wav"
+            ]
+        }
+    },
+    {
+        "id": 35627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008427.wav"
+            ]
+        }
+    },
+    {
+        "id": 35631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006238.wav"
+            ]
+        }
+    },
+    {
+        "id": 35637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005511.wav"
+            ]
+        }
+    },
+    {
+        "id": 35639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000216.wav"
+            ]
+        }
+    },
+    {
+        "id": 35640,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006501.wav"
+            ]
+        }
+    },
+    {
+        "id": 35643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002200.wav"
+            ]
+        }
+    },
+    {
+        "id": 35646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002903.wav"
+            ]
+        }
+    },
+    {
+        "id": 35648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008757.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008757.wav"
+            ]
+        }
+    },
+    {
+        "id": 35650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005213.wav"
+            ]
+        }
+    },
+    {
+        "id": 35653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004567.wav"
+            ]
+        }
+    },
+    {
+        "id": 35658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004639.wav"
+            ]
+        }
+    },
+    {
+        "id": 35659,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000366.wav"
+            ]
+        }
+    },
+    {
+        "id": 35665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002607.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002607.wav"
+            ]
+        }
+    },
+    {
+        "id": 35668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006199.wav"
+            ]
+        }
+    },
+    {
+        "id": 35670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005833.wav"
+            ]
+        }
+    },
+    {
+        "id": 35675,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008536.wav"
+            ]
+        }
+    },
+    {
+        "id": 35686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003355.wav"
+            ]
+        }
+    },
+    {
+        "id": 35688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007105.wav"
+            ]
+        }
+    },
+    {
+        "id": 35697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000639.wav"
+            ]
+        }
+    },
+    {
+        "id": 35700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008732.wav"
+            ]
+        }
+    },
+    {
+        "id": 35705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002297.wav"
+            ]
+        }
+    },
+    {
+        "id": 35710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007828.wav"
+            ]
+        }
+    },
+    {
+        "id": 35713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004388.wav"
+            ]
+        }
+    },
+    {
+        "id": 35721,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003182.wav"
+            ]
+        }
+    },
+    {
+        "id": 35728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005364.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005364.wav"
+            ]
+        }
+    },
+    {
+        "id": 35735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003636.wav"
+            ]
+        }
+    },
+    {
+        "id": 35741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007849.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007849.wav"
+            ]
+        }
+    },
+    {
+        "id": 35744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005304.wav"
+            ]
+        }
+    },
+    {
+        "id": 35750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005351.wav"
+            ]
+        }
+    },
+    {
+        "id": 35752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002765.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002765.wav"
+            ]
+        }
+    },
+    {
+        "id": 35753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002776.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002776.wav"
+            ]
+        }
+    },
+    {
+        "id": 35760,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000392.wav"
+            ]
+        }
+    },
+    {
+        "id": 35761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004687.wav"
+            ]
+        }
+    },
+    {
+        "id": 35763,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006030.wav"
+            ]
+        }
+    },
+    {
+        "id": 35764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003650.wav"
+            ]
+        }
+    },
+    {
+        "id": 35766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008429.wav"
+            ]
+        }
+    },
+    {
+        "id": 35768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008891.wav"
+            ]
+        }
+    },
+    {
+        "id": 35774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006779.wav"
+            ]
+        }
+    },
+    {
+        "id": 35781,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005293.wav"
+            ]
+        }
+    },
+    {
+        "id": 35785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006941.wav"
+            ]
+        }
+    },
+    {
+        "id": 35787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003732.wav"
+            ]
+        }
+    },
+    {
+        "id": 35798,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004519.wav"
+            ]
+        }
+    },
+    {
+        "id": 35799,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007182.wav"
+            ]
+        }
+    },
+    {
+        "id": 35815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006655.wav"
+            ]
+        }
+    },
+    {
+        "id": 35819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007084.wav"
+            ]
+        }
+    },
+    {
+        "id": 35820,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005086.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005086.wav"
+            ]
+        }
+    },
+    {
+        "id": 35829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007618.wav"
+            ]
+        }
+    },
+    {
+        "id": 35850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005588.wav"
+            ]
+        }
+    },
+    {
+        "id": 35851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000855.wav"
+            ]
+        }
+    },
+    {
+        "id": 35852,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003507.wav"
+            ]
+        }
+    },
+    {
+        "id": 35856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005363.wav"
+            ]
+        }
+    },
+    {
+        "id": 35857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002784.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002784.wav"
+            ]
+        }
+    },
+    {
+        "id": 35858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006585.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006585.wav"
+            ]
+        }
+    },
+    {
+        "id": 35860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008601.wav"
+            ]
+        }
+    },
+    {
+        "id": 35862,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004465.wav"
+            ]
+        }
+    },
+    {
+        "id": 35876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006166.wav"
+            ]
+        }
+    },
+    {
+        "id": 35877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002518.wav"
+            ]
+        }
+    },
+    {
+        "id": 35883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004712.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004712.wav"
+            ]
+        }
+    },
+    {
+        "id": 35886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006298.wav"
+            ]
+        }
+    },
+    {
+        "id": 35897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004594.wav"
+            ]
+        }
+    },
+    {
+        "id": 35909,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000256.wav"
+            ]
+        }
+    },
+    {
+        "id": 35917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006151.wav"
+            ]
+        }
+    },
+    {
+        "id": 35925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002577.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002577.wav"
+            ]
+        }
+    },
+    {
+        "id": 35927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002727.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002727.wav"
+            ]
+        }
+    },
+    {
+        "id": 35932,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007519.wav"
+            ]
+        }
+    },
+    {
+        "id": 35933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007778.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007778.wav"
+            ]
+        }
+    },
+    {
+        "id": 35934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007076.wav"
+            ]
+        }
+    },
+    {
+        "id": 35941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005275.wav"
+            ]
+        }
+    },
+    {
+        "id": 35944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001964.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001964.wav"
+            ]
+        }
+    },
+    {
+        "id": 35951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005011.wav"
+            ]
+        }
+    },
+    {
+        "id": 35965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005746.wav"
+            ]
+        }
+    },
+    {
+        "id": 35967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006618.wav"
+            ]
+        }
+    },
+    {
+        "id": 35978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008708.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008708.wav"
+            ]
+        }
+    },
+    {
+        "id": 35986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002392.wav"
+            ]
+        }
+    },
+    {
+        "id": 35999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003293.wav"
+            ]
+        }
+    },
+    {
+        "id": 36001,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003696.wav"
+            ]
+        }
+    },
+    {
+        "id": 36002,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005279.wav"
+            ]
+        }
+    },
+    {
+        "id": 36008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008951.wav"
+            ]
+        }
+    },
+    {
+        "id": 36019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001870.wav"
+            ]
+        }
+    },
+    {
+        "id": 36027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008974.wav"
+            ]
+        }
+    },
+    {
+        "id": 36031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004278.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004278.wav"
+            ]
+        }
+    },
+    {
+        "id": 36056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008724.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008724.wav"
+            ]
+        }
+    },
+    {
+        "id": 36061,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008873.wav"
+            ]
+        }
+    },
+    {
+        "id": 36070,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007077.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007077.wav"
+            ]
+        }
+    },
+    {
+        "id": 36078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003510.wav"
+            ]
+        }
+    },
+    {
+        "id": 36079,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000536.wav"
+            ]
+        }
+    },
+    {
+        "id": 36084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003669.wav"
+            ]
+        }
+    },
+    {
+        "id": 36090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006821.wav"
+            ]
+        }
+    },
+    {
+        "id": 36091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000623.wav"
+            ]
+        }
+    },
+    {
+        "id": 36096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005035.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005035.wav"
+            ]
+        }
+    },
+    {
+        "id": 36097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007655.wav"
+            ]
+        }
+    },
+    {
+        "id": 36107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001409.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001409.wav"
+            ]
+        }
+    },
+    {
+        "id": 36108,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001313.wav"
+            ]
+        }
+    },
+    {
+        "id": 36116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001281.wav"
+            ]
+        }
+    },
+    {
+        "id": 36119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003930.wav"
+            ]
+        }
+    },
+    {
+        "id": 36121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008554.wav"
+            ]
+        }
+    },
+    {
+        "id": 36123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000919.wav"
+            ]
+        }
+    },
+    {
+        "id": 36126,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007900.wav"
+            ]
+        }
+    },
+    {
+        "id": 36132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005438.wav"
+            ]
+        }
+    },
+    {
+        "id": 36135,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich tuba makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000264.wav"
+            ]
+        }
+    },
+    {
+        "id": 36139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006156.wav"
+            ]
+        }
+    },
+    {
+        "id": 36147,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006972.wav"
+            ]
+        }
+    },
+    {
+        "id": 36160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005352.wav"
+            ]
+        }
+    },
+    {
+        "id": 36164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002632.wav"
+            ]
+        }
+    },
+    {
+        "id": 36179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000067.wav"
+            ]
+        }
+    },
+    {
+        "id": 36180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001860.wav"
+            ]
+        }
+    },
+    {
+        "id": 36186,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002560.wav"
+            ]
+        }
+    },
+    {
+        "id": 36189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006632.wav"
+            ]
+        }
+    },
+    {
+        "id": 36193,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006424.wav"
+            ]
+        }
+    },
+    {
+        "id": 36195,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008190.wav"
+            ]
+        }
+    },
+    {
+        "id": 36201,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007373.wav"
+            ]
+        }
+    },
+    {
+        "id": 36208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002561.wav"
+            ]
+        }
+    },
+    {
+        "id": 36209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006822.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006822.wav"
+            ]
+        }
+    },
+    {
+        "id": 36222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005713.wav"
+            ]
+        }
+    },
+    {
+        "id": 36229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007621.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007621.wav"
+            ]
+        }
+    },
+    {
+        "id": 36230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004028.wav"
+            ]
+        }
+    },
+    {
+        "id": 36242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001270.wav"
+            ]
+        }
+    },
+    {
+        "id": 36247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004917.wav"
+            ]
+        }
+    },
+    {
+        "id": 36248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007560.wav"
+            ]
+        }
+    },
+    {
+        "id": 36250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003603.wav"
+            ]
+        }
+    },
+    {
+        "id": 36255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006183.wav"
+            ]
+        }
+    },
+    {
+        "id": 36256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003701.wav"
+            ]
+        }
+    },
+    {
+        "id": 36257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002507.wav"
+            ]
+        }
+    },
+    {
+        "id": 36264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001940.wav"
+            ]
+        }
+    },
+    {
+        "id": 36300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002685.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002685.wav"
+            ]
+        }
+    },
+    {
+        "id": 36307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004525.wav"
+            ]
+        }
+    },
+    {
+        "id": 36321,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000417.wav"
+            ]
+        }
+    },
+    {
+        "id": 36329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004666.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004666.wav"
+            ]
+        }
+    },
+    {
+        "id": 36343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000183.wav"
+            ]
+        }
+    },
+    {
+        "id": 36346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008248.wav"
+            ]
+        }
+    },
+    {
+        "id": 36355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007337.wav"
+            ]
+        }
+    },
+    {
+        "id": 36363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007611.wav"
+            ]
+        }
+    },
+    {
+        "id": 36364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right more rhythmic than the tuba on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002405.wav"
+            ]
+        }
+    },
+    {
+        "id": 36365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003709.wav"
+            ]
+        }
+    },
+    {
+        "id": 36379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003262.wav"
+            ]
+        }
+    },
+    {
+        "id": 36383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006611.wav"
+            ]
+        }
+    },
+    {
+        "id": 36392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004867.wav"
+            ]
+        }
+    },
+    {
+        "id": 36405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002632.wav"
+            ]
+        }
+    },
+    {
+        "id": 36407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008410.wav"
+            ]
+        }
+    },
+    {
+        "id": 36408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007320.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007320.wav"
+            ]
+        }
+    },
+    {
+        "id": 36412,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001129.wav"
+            ]
+        }
+    },
+    {
+        "id": 36424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002933.wav"
+            ]
+        }
+    },
+    {
+        "id": 36425,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005964.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005964.wav"
+            ]
+        }
+    },
+    {
+        "id": 36426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003831.wav"
+            ]
+        }
+    },
+    {
+        "id": 36427,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000062.wav"
+            ]
+        }
+    },
+    {
+        "id": 36432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008156.wav"
+            ]
+        }
+    },
+    {
+        "id": 36436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006541.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006541.wav"
+            ]
+        }
+    },
+    {
+        "id": 36437,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004985.wav"
+            ]
+        }
+    },
+    {
+        "id": 36439,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003704.wav"
+            ]
+        }
+    },
+    {
+        "id": 36440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008670.wav"
+            ]
+        }
+    },
+    {
+        "id": 36444,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002694.wav"
+            ]
+        }
+    },
+    {
+        "id": 36447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008237.wav"
+            ]
+        }
+    },
+    {
+        "id": 36450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000154.wav"
+            ]
+        }
+    },
+    {
+        "id": 36452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006642.wav"
+            ]
+        }
+    },
+    {
+        "id": 36454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004489.wav"
+            ]
+        }
+    },
+    {
+        "id": 36456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006280.wav"
+            ]
+        }
+    },
+    {
+        "id": 36458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006014.wav"
+            ]
+        }
+    },
+    {
+        "id": 36459,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005282.wav"
+            ]
+        }
+    },
+    {
+        "id": 36468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002379.wav"
+            ]
+        }
+    },
+    {
+        "id": 36475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005987.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005987.wav"
+            ]
+        }
+    },
+    {
+        "id": 36486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004732.wav"
+            ]
+        }
+    },
+    {
+        "id": 36487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003906.wav"
+            ]
+        }
+    },
+    {
+        "id": 36495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000650.wav"
+            ]
+        }
+    },
+    {
+        "id": 36503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005509.wav"
+            ]
+        }
+    },
+    {
+        "id": 36504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000084.wav"
+            ]
+        }
+    },
+    {
+        "id": 36506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004285.wav"
+            ]
+        }
+    },
+    {
+        "id": 36507,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000340.wav"
+            ]
+        }
+    },
+    {
+        "id": 36513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000329.wav"
+            ]
+        }
+    },
+    {
+        "id": 36519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004049.wav"
+            ]
+        }
+    },
+    {
+        "id": 36520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008828.wav"
+            ]
+        }
+    },
+    {
+        "id": 36530,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008676.wav"
+            ]
+        }
+    },
+    {
+        "id": 36540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002867.wav"
+            ]
+        }
+    },
+    {
+        "id": 36542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002738.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002738.wav"
+            ]
+        }
+    },
+    {
+        "id": 36548,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004579.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004579.wav"
+            ]
+        }
+    },
+    {
+        "id": 36558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona louder than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003047.wav"
+            ]
+        }
+    },
+    {
+        "id": 36559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005084.wav"
+            ]
+        }
+    },
+    {
+        "id": 36569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich suona makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004729.wav"
+            ]
+        }
+    },
+    {
+        "id": 36575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000435.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000435.wav"
+            ]
+        }
+    },
+    {
+        "id": 36587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001036.wav"
+            ]
+        }
+    },
+    {
+        "id": 36596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004946.wav"
+            ]
+        }
+    },
+    {
+        "id": 36598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007588.wav"
+            ]
+        }
+    },
+    {
+        "id": 36599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003130.wav"
+            ]
+        }
+    },
+    {
+        "id": 36601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008494.wav"
+            ]
+        }
+    },
+    {
+        "id": 36604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006538.wav"
+            ]
+        }
+    },
+    {
+        "id": 36606,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004172.wav"
+            ]
+        }
+    },
+    {
+        "id": 36608,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004216.wav"
+            ]
+        }
+    },
+    {
+        "id": 36609,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002948.wav"
+            ]
+        }
+    },
+    {
+        "id": 36612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006111.wav"
+            ]
+        }
+    },
+    {
+        "id": 36614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007829.wav"
+            ]
+        }
+    },
+    {
+        "id": 36627,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002850.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002850.wav"
+            ]
+        }
+    },
+    {
+        "id": 36628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007656.wav"
+            ]
+        }
+    },
+    {
+        "id": 36629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000505.wav"
+            ]
+        }
+    },
+    {
+        "id": 36631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich xylophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000741.wav"
+            ]
+        }
+    },
+    {
+        "id": 36638,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003445.wav"
+            ]
+        }
+    },
+    {
+        "id": 36667,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007107.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007107.wav"
+            ]
+        }
+    },
+    {
+        "id": 36669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003899.wav"
+            ]
+        }
+    },
+    {
+        "id": 36670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005030.wav"
+            ]
+        }
+    },
+    {
+        "id": 36677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left more rhythmic than the piano on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001194.wav"
+            ]
+        }
+    },
+    {
+        "id": 36684,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003910.wav"
+            ]
+        }
+    },
+    {
+        "id": 36689,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001567.wav"
+            ]
+        }
+    },
+    {
+        "id": 36690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004946.wav"
+            ]
+        }
+    },
+    {
+        "id": 36709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006671.wav"
+            ]
+        }
+    },
+    {
+        "id": 36710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004826.wav"
+            ]
+        }
+    },
+    {
+        "id": 36719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006206.wav"
+            ]
+        }
+    },
+    {
+        "id": 36723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001796.wav"
+            ]
+        }
+    },
+    {
+        "id": 36741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005654.wav"
+            ]
+        }
+    },
+    {
+        "id": 36743,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008975.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008975.wav"
+            ]
+        }
+    },
+    {
+        "id": 36750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001269.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001269.wav"
+            ]
+        }
+    },
+    {
+        "id": 36751,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007283.wav"
+            ]
+        }
+    },
+    {
+        "id": 36757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006420.wav"
+            ]
+        }
+    },
+    {
+        "id": 36758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008769.wav"
+            ]
+        }
+    },
+    {
+        "id": 36767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005122.wav"
+            ]
+        }
+    },
+    {
+        "id": 36769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002833.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002833.wav"
+            ]
+        }
+    },
+    {
+        "id": 36770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002830.wav"
+            ]
+        }
+    },
+    {
+        "id": 36773,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001456.wav"
+            ]
+        }
+    },
+    {
+        "id": 36778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005315.wav"
+            ]
+        }
+    },
+    {
+        "id": 36780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002647.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002647.wav"
+            ]
+        }
+    },
+    {
+        "id": 36782,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005137.wav"
+            ]
+        }
+    },
+    {
+        "id": 36792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002873.wav"
+            ]
+        }
+    },
+    {
+        "id": 36795,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007033.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007033.wav"
+            ]
+        }
+    },
+    {
+        "id": 36797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004609.wav"
+            ]
+        }
+    },
+    {
+        "id": 36802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002798.wav"
+            ]
+        }
+    },
+    {
+        "id": 36807,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002571.wav"
+            ]
+        }
+    },
+    {
+        "id": 36812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007259.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007259.wav"
+            ]
+        }
+    },
+    {
+        "id": 36813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000549.wav"
+            ]
+        }
+    },
+    {
+        "id": 36816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000842.wav"
+            ]
+        }
+    },
+    {
+        "id": 36822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007023.wav"
+            ]
+        }
+    },
+    {
+        "id": 36828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004783.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004783.wav"
+            ]
+        }
+    },
+    {
+        "id": 36841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000460.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000460.wav"
+            ]
+        }
+    },
+    {
+        "id": 36842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003742.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003742.wav"
+            ]
+        }
+    },
+    {
+        "id": 36847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008960.wav"
+            ]
+        }
+    },
+    {
+        "id": 36853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004371.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004371.wav"
+            ]
+        }
+    },
+    {
+        "id": 36859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002296.wav"
+            ]
+        }
+    },
+    {
+        "id": 36873,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005769.wav"
+            ]
+        }
+    },
+    {
+        "id": 36874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005897.wav"
+            ]
+        }
+    },
+    {
+        "id": 36877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000503.wav"
+            ]
+        }
+    },
+    {
+        "id": 36891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007123.wav"
+            ]
+        }
+    },
+    {
+        "id": 36894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002916.wav"
+            ]
+        }
+    },
+    {
+        "id": 36899,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006857.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006857.wav"
+            ]
+        }
+    },
+    {
+        "id": 36903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007701.wav"
+            ]
+        }
+    },
+    {
+        "id": 36905,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008494.wav"
+            ]
+        }
+    },
+    {
+        "id": 36907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006568.wav"
+            ]
+        }
+    },
+    {
+        "id": 36914,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005851.wav"
+            ]
+        }
+    },
+    {
+        "id": 36924,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005267.wav"
+            ]
+        }
+    },
+    {
+        "id": 36929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004762.wav"
+            ]
+        }
+    },
+    {
+        "id": 36930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000554.wav"
+            ]
+        }
+    },
+    {
+        "id": 36943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003414.wav"
+            ]
+        }
+    },
+    {
+        "id": 36952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008665.wav"
+            ]
+        }
+    },
+    {
+        "id": 36953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007108.wav"
+            ]
+        }
+    },
+    {
+        "id": 36962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006240.wav"
+            ]
+        }
+    },
+    {
+        "id": 36973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002153.wav"
+            ]
+        }
+    },
+    {
+        "id": 36985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001517.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001517.wav"
+            ]
+        }
+    },
+    {
+        "id": 36986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000566.wav"
+            ]
+        }
+    },
+    {
+        "id": 36989,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006095.wav"
+            ]
+        }
+    },
+    {
+        "id": 36990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003401.wav"
+            ]
+        }
+    },
+    {
+        "id": 36997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002303.wav"
+            ]
+        }
+    },
+    {
+        "id": 36998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004258.wav"
+            ]
+        }
+    },
+    {
+        "id": 37000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008048.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008048.wav"
+            ]
+        }
+    },
+    {
+        "id": 37007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006046.wav"
+            ]
+        }
+    },
+    {
+        "id": 37008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008292.wav"
+            ]
+        }
+    },
+    {
+        "id": 37011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003845.wav"
+            ]
+        }
+    },
+    {
+        "id": 37029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002689.wav"
+            ]
+        }
+    },
+    {
+        "id": 37031,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006951.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006951.wav"
+            ]
+        }
+    },
+    {
+        "id": 37033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005046.wav"
+            ]
+        }
+    },
+    {
+        "id": 37039,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006249.wav"
+            ]
+        }
+    },
+    {
+        "id": 37040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008783.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008783.wav"
+            ]
+        }
+    },
+    {
+        "id": 37043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004863.wav"
+            ]
+        }
+    },
+    {
+        "id": 37081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004542.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004542.wav"
+            ]
+        }
+    },
+    {
+        "id": 37089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005453.wav"
+            ]
+        }
+    },
+    {
+        "id": 37091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008379.wav"
+            ]
+        }
+    },
+    {
+        "id": 37092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004103.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004103.wav"
+            ]
+        }
+    },
+    {
+        "id": 37098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007405.wav"
+            ]
+        }
+    },
+    {
+        "id": 37102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004645.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004645.wav"
+            ]
+        }
+    },
+    {
+        "id": 37109,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007825.wav"
+            ]
+        }
+    },
+    {
+        "id": 37113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007589.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007589.wav"
+            ]
+        }
+    },
+    {
+        "id": 37114,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003943.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003943.wav"
+            ]
+        }
+    },
+    {
+        "id": 37118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007703.wav"
+            ]
+        }
+    },
+    {
+        "id": 37125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000932.wav"
+            ]
+        }
+    },
+    {
+        "id": 37127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007529.wav"
+            ]
+        }
+    },
+    {
+        "id": 37128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006458.wav"
+            ]
+        }
+    },
+    {
+        "id": 37132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003388.wav"
+            ]
+        }
+    },
+    {
+        "id": 37145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005148.wav"
+            ]
+        }
+    },
+    {
+        "id": 37160,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002438.wav"
+            ]
+        }
+    },
+    {
+        "id": 37164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006984.wav"
+            ]
+        }
+    },
+    {
+        "id": 37166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008693.wav"
+            ]
+        }
+    },
+    {
+        "id": 37168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002075.wav"
+            ]
+        }
+    },
+    {
+        "id": 37173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006420.wav"
+            ]
+        }
+    },
+    {
+        "id": 37175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008976.wav"
+            ]
+        }
+    },
+    {
+        "id": 37194,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007728.wav"
+            ]
+        }
+    },
+    {
+        "id": 37206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005507.wav"
+            ]
+        }
+    },
+    {
+        "id": 37210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006690.wav"
+            ]
+        }
+    },
+    {
+        "id": 37214,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005775.wav"
+            ]
+        }
+    },
+    {
+        "id": 37216,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005222.wav"
+            ]
+        }
+    },
+    {
+        "id": 37226,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008070.wav"
+            ]
+        }
+    },
+    {
+        "id": 37259,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006030.wav"
+            ]
+        }
+    },
+    {
+        "id": 37261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003082.wav"
+            ]
+        }
+    },
+    {
+        "id": 37271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005115.wav"
+            ]
+        }
+    },
+    {
+        "id": 37274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005520.wav"
+            ]
+        }
+    },
+    {
+        "id": 37281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004581.wav"
+            ]
+        }
+    },
+    {
+        "id": 37299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001046.wav"
+            ]
+        }
+    },
+    {
+        "id": 37300,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004130.wav"
+            ]
+        }
+    },
+    {
+        "id": 37307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008949.wav"
+            ]
+        }
+    },
+    {
+        "id": 37316,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002180.wav"
+            ]
+        }
+    },
+    {
+        "id": 37317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001147.wav"
+            ]
+        }
+    },
+    {
+        "id": 37325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008661.wav"
+            ]
+        }
+    },
+    {
+        "id": 37339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008628.wav"
+            ]
+        }
+    },
+    {
+        "id": 37340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003476.wav"
+            ]
+        }
+    },
+    {
+        "id": 37346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005037.wav"
+            ]
+        }
+    },
+    {
+        "id": 37351,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003804.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003804.wav"
+            ]
+        }
+    },
+    {
+        "id": 37373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005981.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005981.wav"
+            ]
+        }
+    },
+    {
+        "id": 37374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001052.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001052.wav"
+            ]
+        }
+    },
+    {
+        "id": 37376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006558.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006558.wav"
+            ]
+        }
+    },
+    {
+        "id": 37377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007144.wav"
+            ]
+        }
+    },
+    {
+        "id": 37381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005907.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005907.wav"
+            ]
+        }
+    },
+    {
+        "id": 37395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005070.wav"
+            ]
+        }
+    },
+    {
+        "id": 37404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008775.wav"
+            ]
+        }
+    },
+    {
+        "id": 37408,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001600.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001600.wav"
+            ]
+        }
+    },
+    {
+        "id": 37413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006818.wav"
+            ]
+        }
+    },
+    {
+        "id": 37416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007676.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007676.wav"
+            ]
+        }
+    },
+    {
+        "id": 37419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008431.wav"
+            ]
+        }
+    },
+    {
+        "id": 37423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004142.wav"
+            ]
+        }
+    },
+    {
+        "id": 37430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000587.wav"
+            ]
+        }
+    },
+    {
+        "id": 37435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003825.wav"
+            ]
+        }
+    },
+    {
+        "id": 37436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004928.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004928.wav"
+            ]
+        }
+    },
+    {
+        "id": 37448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004187.wav"
+            ]
+        }
+    },
+    {
+        "id": 37450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003130.wav"
+            ]
+        }
+    },
+    {
+        "id": 37454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005529.wav"
+            ]
+        }
+    },
+    {
+        "id": 37465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001156.wav"
+            ]
+        }
+    },
+    {
+        "id": 37471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000304.wav"
+            ]
+        }
+    },
+    {
+        "id": 37479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003441.wav"
+            ]
+        }
+    },
+    {
+        "id": 37483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000490.wav"
+            ]
+        }
+    },
+    {
+        "id": 37485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000323.wav"
+            ]
+        }
+    },
+    {
+        "id": 37490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004108.wav"
+            ]
+        }
+    },
+    {
+        "id": 37492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008323.wav"
+            ]
+        }
+    },
+    {
+        "id": 37494,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005773.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005773.wav"
+            ]
+        }
+    },
+    {
+        "id": 37496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008650.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008650.wav"
+            ]
+        }
+    },
+    {
+        "id": 37498,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006768.wav"
+            ]
+        }
+    },
+    {
+        "id": 37502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002632.wav"
+            ]
+        }
+    },
+    {
+        "id": 37507,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002827.wav"
+            ]
+        }
+    },
+    {
+        "id": 37508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002104.wav"
+            ]
+        }
+    },
+    {
+        "id": 37516,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002490.wav"
+            ]
+        }
+    },
+    {
+        "id": 37545,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000394.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000394.wav"
+            ]
+        }
+    },
+    {
+        "id": 37547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003134.wav"
+            ]
+        }
+    },
+    {
+        "id": 37552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007535.wav"
+            ]
+        }
+    },
+    {
+        "id": 37554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003132.wav"
+            ]
+        }
+    },
+    {
+        "id": 37557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005533.wav"
+            ]
+        }
+    },
+    {
+        "id": 37559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004707.wav"
+            ]
+        }
+    },
+    {
+        "id": 37560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005664.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005664.wav"
+            ]
+        }
+    },
+    {
+        "id": 37568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002429.wav"
+            ]
+        }
+    },
+    {
+        "id": 37574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004002.wav"
+            ]
+        }
+    },
+    {
+        "id": 37575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005701.wav"
+            ]
+        }
+    },
+    {
+        "id": 37590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006599.wav"
+            ]
+        }
+    },
+    {
+        "id": 37595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000550.wav"
+            ]
+        }
+    },
+    {
+        "id": 37607,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006206.wav"
+            ]
+        }
+    },
+    {
+        "id": 37614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001826.wav"
+            ]
+        }
+    },
+    {
+        "id": 37624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008134.wav"
+            ]
+        }
+    },
+    {
+        "id": 37625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008917.wav"
+            ]
+        }
+    },
+    {
+        "id": 37631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007286.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007286.wav"
+            ]
+        }
+    },
+    {
+        "id": 37632,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002813.wav"
+            ]
+        }
+    },
+    {
+        "id": 37648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006093.wav"
+            ]
+        }
+    },
+    {
+        "id": 37653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008639.wav"
+            ]
+        }
+    },
+    {
+        "id": 37660,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006780.wav"
+            ]
+        }
+    },
+    {
+        "id": 37662,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008838.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008838.wav"
+            ]
+        }
+    },
+    {
+        "id": 37670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006705.wav"
+            ]
+        }
+    },
+    {
+        "id": 37677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005301.wav"
+            ]
+        }
+    },
+    {
+        "id": 37691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008010.wav"
+            ]
+        }
+    },
+    {
+        "id": 37693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000756.wav"
+            ]
+        }
+    },
+    {
+        "id": 37701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007690.wav"
+            ]
+        }
+    },
+    {
+        "id": 37705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006764.wav"
+            ]
+        }
+    },
+    {
+        "id": 37710,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004557.wav"
+            ]
+        }
+    },
+    {
+        "id": 37711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008226.wav"
+            ]
+        }
+    },
+    {
+        "id": 37712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003076.wav"
+            ]
+        }
+    },
+    {
+        "id": 37718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007256.wav"
+            ]
+        }
+    },
+    {
+        "id": 37723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007983.wav"
+            ]
+        }
+    },
+    {
+        "id": 37727,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005914.wav"
+            ]
+        }
+    },
+    {
+        "id": 37733,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002329.wav"
+            ]
+        }
+    },
+    {
+        "id": 37736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005560.wav"
+            ]
+        }
+    },
+    {
+        "id": 37755,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002987.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002987.wav"
+            ]
+        }
+    },
+    {
+        "id": 37757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004397.wav"
+            ]
+        }
+    },
+    {
+        "id": 37758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000575.wav"
+            ]
+        }
+    },
+    {
+        "id": 37767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002378.wav"
+            ]
+        }
+    },
+    {
+        "id": 37772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002597.wav"
+            ]
+        }
+    },
+    {
+        "id": 37774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000396.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000396.wav"
+            ]
+        }
+    },
+    {
+        "id": 37777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006509.wav"
+            ]
+        }
+    },
+    {
+        "id": 37778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003756.wav"
+            ]
+        }
+    },
+    {
+        "id": 37783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008807.wav"
+            ]
+        }
+    },
+    {
+        "id": 37785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002430.wav"
+            ]
+        }
+    },
+    {
+        "id": 37787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002933.wav"
+            ]
+        }
+    },
+    {
+        "id": 37792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004014.wav"
+            ]
+        }
+    },
+    {
+        "id": 37796,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003527.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003527.wav"
+            ]
+        }
+    },
+    {
+        "id": 37800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003244.wav"
+            ]
+        }
+    },
+    {
+        "id": 37814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007974.wav"
+            ]
+        }
+    },
+    {
+        "id": 37824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008909.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008909.wav"
+            ]
+        }
+    },
+    {
+        "id": 37825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002590.wav"
+            ]
+        }
+    },
+    {
+        "id": 37828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006227.wav"
+            ]
+        }
+    },
+    {
+        "id": 37835,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002503.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002503.wav"
+            ]
+        }
+    },
+    {
+        "id": 37843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005134.wav"
+            ]
+        }
+    },
+    {
+        "id": 37846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006006.wav"
+            ]
+        }
+    },
+    {
+        "id": 37851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007429.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007429.wav"
+            ]
+        }
+    },
+    {
+        "id": 37854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele playing longer than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000580.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000580.wav"
+            ]
+        }
+    },
+    {
+        "id": 37864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005180.wav"
+            ]
+        }
+    },
+    {
+        "id": 37866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005073.wav"
+            ]
+        }
+    },
+    {
+        "id": 37869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007138.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007138.wav"
+            ]
+        }
+    },
+    {
+        "id": 37872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003887.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003887.wav"
+            ]
+        }
+    },
+    {
+        "id": 37874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000567.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000567.wav"
+            ]
+        }
+    },
+    {
+        "id": 37876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001031.wav"
+            ]
+        }
+    },
+    {
+        "id": 37877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007237.wav"
+            ]
+        }
+    },
+    {
+        "id": 37878,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002889.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002889.wav"
+            ]
+        }
+    },
+    {
+        "id": 37882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007041.wav"
+            ]
+        }
+    },
+    {
+        "id": 37883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004596.wav"
+            ]
+        }
+    },
+    {
+        "id": 37885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006381.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006381.wav"
+            ]
+        }
+    },
+    {
+        "id": 37889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000378.wav"
+            ]
+        }
+    },
+    {
+        "id": 37890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008445.wav"
+            ]
+        }
+    },
+    {
+        "id": 37891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007885.wav"
+            ]
+        }
+    },
+    {
+        "id": 37899,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005715.wav"
+            ]
+        }
+    },
+    {
+        "id": 37900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008693.wav"
+            ]
+        }
+    },
+    {
+        "id": 37901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000200.wav"
+            ]
+        }
+    },
+    {
+        "id": 37902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left louder than the congas on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001441.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001441.wav"
+            ]
+        }
+    },
+    {
+        "id": 37905,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004547.wav"
+            ]
+        }
+    },
+    {
+        "id": 37907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005901.wav"
+            ]
+        }
+    },
+    {
+        "id": 37908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004969.wav"
+            ]
+        }
+    },
+    {
+        "id": 37916,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008432.wav"
+            ]
+        }
+    },
+    {
+        "id": 37918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006689.wav"
+            ]
+        }
+    },
+    {
+        "id": 37927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008725.wav"
+            ]
+        }
+    },
+    {
+        "id": 37930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008343.wav"
+            ]
+        }
+    },
+    {
+        "id": 37931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008652.wav"
+            ]
+        }
+    },
+    {
+        "id": 37933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006674.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006674.wav"
+            ]
+        }
+    },
+    {
+        "id": 37936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002711.wav"
+            ]
+        }
+    },
+    {
+        "id": 37939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005462.wav"
+            ]
+        }
+    },
+    {
+        "id": 37944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006550.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006550.wav"
+            ]
+        }
+    },
+    {
+        "id": 37946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002173.wav"
+            ]
+        }
+    },
+    {
+        "id": 37947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006237.wav"
+            ]
+        }
+    },
+    {
+        "id": 37961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005379.wav"
+            ]
+        }
+    },
+    {
+        "id": 37964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001693.wav"
+            ]
+        }
+    },
+    {
+        "id": 37973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001739.wav"
+            ]
+        }
+    },
+    {
+        "id": 37981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004254.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004254.wav"
+            ]
+        }
+    },
+    {
+        "id": 37986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004467.wav"
+            ]
+        }
+    },
+    {
+        "id": 37991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002149.wav"
+            ]
+        }
+    },
+    {
+        "id": 37992,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000087.wav"
+            ]
+        }
+    },
+    {
+        "id": 37997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005393.wav"
+            ]
+        }
+    },
+    {
+        "id": 37998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004516.wav"
+            ]
+        }
+    },
+    {
+        "id": 38007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007665.wav"
+            ]
+        }
+    },
+    {
+        "id": 38008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001098.wav"
+            ]
+        }
+    },
+    {
+        "id": 38018,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007580.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007580.wav"
+            ]
+        }
+    },
+    {
+        "id": 38027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001324.wav"
+            ]
+        }
+    },
+    {
+        "id": 38033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004210.wav"
+            ]
+        }
+    },
+    {
+        "id": 38041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008563.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008563.wav"
+            ]
+        }
+    },
+    {
+        "id": 38056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000194.wav"
+            ]
+        }
+    },
+    {
+        "id": 38066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006317.wav"
+            ]
+        }
+    },
+    {
+        "id": 38068,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006712.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006712.wav"
+            ]
+        }
+    },
+    {
+        "id": 38073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005910.wav"
+            ]
+        }
+    },
+    {
+        "id": 38077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002183.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002183.wav"
+            ]
+        }
+    },
+    {
+        "id": 38098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004236.wav"
+            ]
+        }
+    },
+    {
+        "id": 38099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005161.wav"
+            ]
+        }
+    },
+    {
+        "id": 38101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008898.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008898.wav"
+            ]
+        }
+    },
+    {
+        "id": 38104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002490.wav"
+            ]
+        }
+    },
+    {
+        "id": 38115,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008287.wav"
+            ]
+        }
+    },
+    {
+        "id": 38119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008404.wav"
+            ]
+        }
+    },
+    {
+        "id": 38121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004584.wav"
+            ]
+        }
+    },
+    {
+        "id": 38122,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008059.wav"
+            ]
+        }
+    },
+    {
+        "id": 38124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004241.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004241.wav"
+            ]
+        }
+    },
+    {
+        "id": 38129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004100.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004100.wav"
+            ]
+        }
+    },
+    {
+        "id": 38132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007920.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007920.wav"
+            ]
+        }
+    },
+    {
+        "id": 38133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008622.wav"
+            ]
+        }
+    },
+    {
+        "id": 38136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000556.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000556.wav"
+            ]
+        }
+    },
+    {
+        "id": 38137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000791.wav"
+            ]
+        }
+    },
+    {
+        "id": 38141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001367.wav"
+            ]
+        }
+    },
+    {
+        "id": 38149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007155.wav"
+            ]
+        }
+    },
+    {
+        "id": 38150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many musical instruments were heard throughout the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000512.wav"
+            ]
+        }
+    },
+    {
+        "id": 38159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003070.wav"
+            ]
+        }
+    },
+    {
+        "id": 38162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006262.wav"
+            ]
+        }
+    },
+    {
+        "id": 38166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006133.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006133.wav"
+            ]
+        }
+    },
+    {
+        "id": 38170,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007196.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007196.wav"
+            ]
+        }
+    },
+    {
+        "id": 38178,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003574.wav"
+            ]
+        }
+    },
+    {
+        "id": 38179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004952.wav"
+            ]
+        }
+    },
+    {
+        "id": 38184,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007752.wav"
+            ]
+        }
+    },
+    {
+        "id": 38186,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001492.wav"
+            ]
+        }
+    },
+    {
+        "id": 38191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007019.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007019.wav"
+            ]
+        }
+    },
+    {
+        "id": 38197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003932.wav"
+            ]
+        }
+    },
+    {
+        "id": 38198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004957.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004957.wav"
+            ]
+        }
+    },
+    {
+        "id": 38199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004427.wav"
+            ]
+        }
+    },
+    {
+        "id": 38206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007149.wav"
+            ]
+        }
+    },
+    {
+        "id": 38210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001267.wav"
+            ]
+        }
+    },
+    {
+        "id": 38211,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000259.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000259.wav"
+            ]
+        }
+    },
+    {
+        "id": 38217,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005626.wav"
+            ]
+        }
+    },
+    {
+        "id": 38218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002040.wav"
+            ]
+        }
+    },
+    {
+        "id": 38219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006129.wav"
+            ]
+        }
+    },
+    {
+        "id": 38222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003047.wav"
+            ]
+        }
+    },
+    {
+        "id": 38228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008151.wav"
+            ]
+        }
+    },
+    {
+        "id": 38231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007144.wav"
+            ]
+        }
+    },
+    {
+        "id": 38232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008332.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008332.wav"
+            ]
+        }
+    },
+    {
+        "id": 38238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005041.wav"
+            ]
+        }
+    },
+    {
+        "id": 38243,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002547.wav"
+            ]
+        }
+    },
+    {
+        "id": 38258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005243.wav"
+            ]
+        }
+    },
+    {
+        "id": 38260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006899.wav"
+            ]
+        }
+    },
+    {
+        "id": 38263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008316.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008316.wav"
+            ]
+        }
+    },
+    {
+        "id": 38266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000179.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000179.wav"
+            ]
+        }
+    },
+    {
+        "id": 38275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004270.wav"
+            ]
+        }
+    },
+    {
+        "id": 38293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000190.wav"
+            ]
+        }
+    },
+    {
+        "id": 38301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003943.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003943.wav"
+            ]
+        }
+    },
+    {
+        "id": 38302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006187.wav"
+            ]
+        }
+    },
+    {
+        "id": 38306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005166.wav"
+            ]
+        }
+    },
+    {
+        "id": 38308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008358.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008358.wav"
+            ]
+        }
+    },
+    {
+        "id": 38311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001840.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001840.wav"
+            ]
+        }
+    },
+    {
+        "id": 38315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003482.wav"
+            ]
+        }
+    },
+    {
+        "id": 38316,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001129.wav"
+            ]
+        }
+    },
+    {
+        "id": 38325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001354.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001354.wav"
+            ]
+        }
+    },
+    {
+        "id": 38340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000694.wav"
+            ]
+        }
+    },
+    {
+        "id": 38341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008751.wav"
+            ]
+        }
+    },
+    {
+        "id": 38343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007782.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007782.wav"
+            ]
+        }
+    },
+    {
+        "id": 38357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007582.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007582.wav"
+            ]
+        }
+    },
+    {
+        "id": 38363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003401.wav"
+            ]
+        }
+    },
+    {
+        "id": 38364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006959.wav"
+            ]
+        }
+    },
+    {
+        "id": 38370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003982.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003982.wav"
+            ]
+        }
+    },
+    {
+        "id": 38372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007741.wav"
+            ]
+        }
+    },
+    {
+        "id": 38389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004223.wav"
+            ]
+        }
+    },
+    {
+        "id": 38391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007741.wav"
+            ]
+        }
+    },
+    {
+        "id": 38407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003102.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003102.wav"
+            ]
+        }
+    },
+    {
+        "id": 38411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007060.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007060.wav"
+            ]
+        }
+    },
+    {
+        "id": 38415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001003.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001003.wav"
+            ]
+        }
+    },
+    {
+        "id": 38419,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007785.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007785.wav"
+            ]
+        }
+    },
+    {
+        "id": 38420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006728.wav"
+            ]
+        }
+    },
+    {
+        "id": 38424,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007967.wav"
+            ]
+        }
+    },
+    {
+        "id": 38430,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000298.wav"
+            ]
+        }
+    },
+    {
+        "id": 38434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004054.wav"
+            ]
+        }
+    },
+    {
+        "id": 38446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003835.wav"
+            ]
+        }
+    },
+    {
+        "id": 38449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006586.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006586.wav"
+            ]
+        }
+    },
+    {
+        "id": 38455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich xylophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000608.wav"
+            ]
+        }
+    },
+    {
+        "id": 38461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005573.wav"
+            ]
+        }
+    },
+    {
+        "id": 38462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000351.wav"
+            ]
+        }
+    },
+    {
+        "id": 38466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004886.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004886.wav"
+            ]
+        }
+    },
+    {
+        "id": 38471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000608.wav"
+            ]
+        }
+    },
+    {
+        "id": 38485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005118.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005118.wav"
+            ]
+        }
+    },
+    {
+        "id": 38486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000960.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000960.wav"
+            ]
+        }
+    },
+    {
+        "id": 38487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004238.wav"
+            ]
+        }
+    },
+    {
+        "id": 38490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006063.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006063.wav"
+            ]
+        }
+    },
+    {
+        "id": 38491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005399.wav"
+            ]
+        }
+    },
+    {
+        "id": 38493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003516.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003516.wav"
+            ]
+        }
+    },
+    {
+        "id": 38496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007904.wav"
+            ]
+        }
+    },
+    {
+        "id": 38497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002581.wav"
+            ]
+        }
+    },
+    {
+        "id": 38511,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006569.wav"
+            ]
+        }
+    },
+    {
+        "id": 38513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002590.wav"
+            ]
+        }
+    },
+    {
+        "id": 38514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008593.wav"
+            ]
+        }
+    },
+    {
+        "id": 38515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005301.wav"
+            ]
+        }
+    },
+    {
+        "id": 38529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000810.wav"
+            ]
+        }
+    },
+    {
+        "id": 38532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008524.wav"
+            ]
+        }
+    },
+    {
+        "id": 38535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000895.wav"
+            ]
+        }
+    },
+    {
+        "id": 38536,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006824.wav"
+            ]
+        }
+    },
+    {
+        "id": 38540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005904.wav"
+            ]
+        }
+    },
+    {
+        "id": 38549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005042.wav"
+            ]
+        }
+    },
+    {
+        "id": 38552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007463.wav"
+            ]
+        }
+    },
+    {
+        "id": 38554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003379.wav"
+            ]
+        }
+    },
+    {
+        "id": 38557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008541.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008541.wav"
+            ]
+        }
+    },
+    {
+        "id": 38560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006505.wav"
+            ]
+        }
+    },
+    {
+        "id": 38566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004304.wav"
+            ]
+        }
+    },
+    {
+        "id": 38568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006873.wav"
+            ]
+        }
+    },
+    {
+        "id": 38572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000194.wav"
+            ]
+        }
+    },
+    {
+        "id": 38587,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000882.wav"
+            ]
+        }
+    },
+    {
+        "id": 38593,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008158.wav"
+            ]
+        }
+    },
+    {
+        "id": 38595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000412.wav"
+            ]
+        }
+    },
+    {
+        "id": 38596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005624.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005624.wav"
+            ]
+        }
+    },
+    {
+        "id": 38616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004436.wav"
+            ]
+        }
+    },
+    {
+        "id": 38621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004059.wav"
+            ]
+        }
+    },
+    {
+        "id": 38622,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005250.wav"
+            ]
+        }
+    },
+    {
+        "id": 38625,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003088.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003088.wav"
+            ]
+        }
+    },
+    {
+        "id": 38630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001965.wav"
+            ]
+        }
+    },
+    {
+        "id": 38647,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006785.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006785.wav"
+            ]
+        }
+    },
+    {
+        "id": 38655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004157.wav"
+            ]
+        }
+    },
+    {
+        "id": 38658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004701.wav"
+            ]
+        }
+    },
+    {
+        "id": 38663,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008663.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008663.wav"
+            ]
+        }
+    },
+    {
+        "id": 38666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000298.wav"
+            ]
+        }
+    },
+    {
+        "id": 38670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005718.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005718.wav"
+            ]
+        }
+    },
+    {
+        "id": 38675,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006318.wav"
+            ]
+        }
+    },
+    {
+        "id": 38693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003531.wav"
+            ]
+        }
+    },
+    {
+        "id": 38700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006221.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006221.wav"
+            ]
+        }
+    },
+    {
+        "id": 38712,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002446.wav"
+            ]
+        }
+    },
+    {
+        "id": 38716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002642.wav"
+            ]
+        }
+    },
+    {
+        "id": 38725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006635.wav"
+            ]
+        }
+    },
+    {
+        "id": 38734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004659.wav"
+            ]
+        }
+    },
+    {
+        "id": 38737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004395.wav"
+            ]
+        }
+    },
+    {
+        "id": 38752,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007540.wav"
+            ]
+        }
+    },
+    {
+        "id": 38754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004326.wav"
+            ]
+        }
+    },
+    {
+        "id": 38756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003093.wav"
+            ]
+        }
+    },
+    {
+        "id": 38761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003340.wav"
+            ]
+        }
+    },
+    {
+        "id": 38767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005301.wav"
+            ]
+        }
+    },
+    {
+        "id": 38770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005314.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005314.wav"
+            ]
+        }
+    },
+    {
+        "id": 38772,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008672.wav"
+            ]
+        }
+    },
+    {
+        "id": 38778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002942.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002942.wav"
+            ]
+        }
+    },
+    {
+        "id": 38784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003934.wav"
+            ]
+        }
+    },
+    {
+        "id": 38789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004126.wav"
+            ]
+        }
+    },
+    {
+        "id": 38794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005165.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005165.wav"
+            ]
+        }
+    },
+    {
+        "id": 38797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007549.wav"
+            ]
+        }
+    },
+    {
+        "id": 38804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone on the left louder than the xylophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003184.wav"
+            ]
+        }
+    },
+    {
+        "id": 38809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005266.wav"
+            ]
+        }
+    },
+    {
+        "id": 38816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000422.wav"
+            ]
+        }
+    },
+    {
+        "id": 38818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00008963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00008963.wav"
+            ]
+        }
+    },
+    {
+        "id": 38823,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004671.wav"
+            ]
+        }
+    },
+    {
+        "id": 38825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002391.wav"
+            ]
+        }
+    },
+    {
+        "id": 38830,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006918.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006918.wav"
+            ]
+        }
+    },
+    {
+        "id": 38834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007716.wav"
+            ]
+        }
+    },
+    {
+        "id": 38843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007484.wav"
+            ]
+        }
+    },
+    {
+        "id": 38844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001952.wav"
+            ]
+        }
+    },
+    {
+        "id": 38845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006362.wav"
+            ]
+        }
+    },
+    {
+        "id": 38855,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003159.wav"
+            ]
+        }
+    },
+    {
+        "id": 38856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003059.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003059.wav"
+            ]
+        }
+    },
+    {
+        "id": 38859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003078.wav"
+            ]
+        }
+    },
+    {
+        "id": 38861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003293.wav"
+            ]
+        }
+    },
+    {
+        "id": 38865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000365.wav"
+            ]
+        }
+    },
+    {
+        "id": 38866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004438.wav"
+            ]
+        }
+    },
+    {
+        "id": 38887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008554.wav"
+            ]
+        }
+    },
+    {
+        "id": 38890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003654.wav"
+            ]
+        }
+    },
+    {
+        "id": 38907,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006462.wav"
+            ]
+        }
+    },
+    {
+        "id": 38908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003494.wav"
+            ]
+        }
+    },
+    {
+        "id": 38911,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000547.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000547.wav"
+            ]
+        }
+    },
+    {
+        "id": 38921,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003758.wav"
+            ]
+        }
+    },
+    {
+        "id": 38930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002755.wav"
+            ]
+        }
+    },
+    {
+        "id": 38934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007432.wav"
+            ]
+        }
+    },
+    {
+        "id": 38946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008534.wav"
+            ]
+        }
+    },
+    {
+        "id": 38955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008116.wav"
+            ]
+        }
+    },
+    {
+        "id": 38959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005518.wav"
+            ]
+        }
+    },
+    {
+        "id": 38962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003093.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003093.wav"
+            ]
+        }
+    },
+    {
+        "id": 38971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007386.wav"
+            ]
+        }
+    },
+    {
+        "id": 38973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004559.wav"
+            ]
+        }
+    },
+    {
+        "id": 38978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000455.wav"
+            ]
+        }
+    },
+    {
+        "id": 38983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003271.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003271.wav"
+            ]
+        }
+    },
+    {
+        "id": 38993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006055.wav"
+            ]
+        }
+    },
+    {
+        "id": 38994,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003983.wav"
+            ]
+        }
+    },
+    {
+        "id": 39000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008420.wav"
+            ]
+        }
+    },
+    {
+        "id": 39004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the left louder than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004809.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004809.wav"
+            ]
+        }
+    },
+    {
+        "id": 39005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005391.wav"
+            ]
+        }
+    },
+    {
+        "id": 39011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000554.wav"
+            ]
+        }
+    },
+    {
+        "id": 39013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003270.wav"
+            ]
+        }
+    },
+    {
+        "id": 39021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002978.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002978.wav"
+            ]
+        }
+    },
+    {
+        "id": 39028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000505.wav"
+            ]
+        }
+    },
+    {
+        "id": 39029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008775.wav"
+            ]
+        }
+    },
+    {
+        "id": 39040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000375.wav"
+            ]
+        }
+    },
+    {
+        "id": 39041,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005426.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005426.wav"
+            ]
+        }
+    },
+    {
+        "id": 39046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001878.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001878.wav"
+            ]
+        }
+    },
+    {
+        "id": 39049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000931.wav"
+            ]
+        }
+    },
+    {
+        "id": 39055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005117.wav"
+            ]
+        }
+    },
+    {
+        "id": 39057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008498.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008498.wav"
+            ]
+        }
+    },
+    {
+        "id": 39058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004284.wav"
+            ]
+        }
+    },
+    {
+        "id": 39064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003883.wav"
+            ]
+        }
+    },
+    {
+        "id": 39065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006260.wav"
+            ]
+        }
+    },
+    {
+        "id": 39068,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002852.wav"
+            ]
+        }
+    },
+    {
+        "id": 39077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004682.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004682.wav"
+            ]
+        }
+    },
+    {
+        "id": 39080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002785.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002785.wav"
+            ]
+        }
+    },
+    {
+        "id": 39082,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006529.wav"
+            ]
+        }
+    },
+    {
+        "id": 39084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003747.wav"
+            ]
+        }
+    },
+    {
+        "id": 39087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003702.wav"
+            ]
+        }
+    },
+    {
+        "id": 39094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005294.wav"
+            ]
+        }
+    },
+    {
+        "id": 39096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004281.wav"
+            ]
+        }
+    },
+    {
+        "id": 39101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002855.wav"
+            ]
+        }
+    },
+    {
+        "id": 39103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002679.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002679.wav"
+            ]
+        }
+    },
+    {
+        "id": 39117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005758.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005758.wav"
+            ]
+        }
+    },
+    {
+        "id": 39127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008163.wav"
+            ]
+        }
+    },
+    {
+        "id": 39132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004002.wav"
+            ]
+        }
+    },
+    {
+        "id": 39136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004799.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004799.wav"
+            ]
+        }
+    },
+    {
+        "id": 39137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004971.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004971.wav"
+            ]
+        }
+    },
+    {
+        "id": 39147,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000098.wav"
+            ]
+        }
+    },
+    {
+        "id": 39148,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007333.wav"
+            ]
+        }
+    },
+    {
+        "id": 39149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008492.wav"
+            ]
+        }
+    },
+    {
+        "id": 39150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000011.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000011.wav"
+            ]
+        }
+    },
+    {
+        "id": 39153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008880.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008880.wav"
+            ]
+        }
+    },
+    {
+        "id": 39167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003379.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003379.wav"
+            ]
+        }
+    },
+    {
+        "id": 39168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008877.wav"
+            ]
+        }
+    },
+    {
+        "id": 39169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005827.wav"
+            ]
+        }
+    },
+    {
+        "id": 39175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006682.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006682.wav"
+            ]
+        }
+    },
+    {
+        "id": 39179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005759.wav"
+            ]
+        }
+    },
+    {
+        "id": 39208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000346.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000346.wav"
+            ]
+        }
+    },
+    {
+        "id": 39209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006244.wav"
+            ]
+        }
+    },
+    {
+        "id": 39210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004507.wav"
+            ]
+        }
+    },
+    {
+        "id": 39232,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008652.wav"
+            ]
+        }
+    },
+    {
+        "id": 39237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007233.wav"
+            ]
+        }
+    },
+    {
+        "id": 39238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003340.wav"
+            ]
+        }
+    },
+    {
+        "id": 39244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007064.wav"
+            ]
+        }
+    },
+    {
+        "id": 39254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004338.wav"
+            ]
+        }
+    },
+    {
+        "id": 39263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006615.wav"
+            ]
+        }
+    },
+    {
+        "id": 39275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008743.wav"
+            ]
+        }
+    },
+    {
+        "id": 39286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006731.wav"
+            ]
+        }
+    },
+    {
+        "id": 39289,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003554.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003554.wav"
+            ]
+        }
+    },
+    {
+        "id": 39294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005116.wav"
+            ]
+        }
+    },
+    {
+        "id": 39299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005313.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005313.wav"
+            ]
+        }
+    },
+    {
+        "id": 39306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001754.wav"
+            ]
+        }
+    },
+    {
+        "id": 39308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005165.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005165.wav"
+            ]
+        }
+    },
+    {
+        "id": 39310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008030.wav"
+            ]
+        }
+    },
+    {
+        "id": 39314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007638.wav"
+            ]
+        }
+    },
+    {
+        "id": 39318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003444.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003444.wav"
+            ]
+        }
+    },
+    {
+        "id": 39324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000638.wav"
+            ]
+        }
+    },
+    {
+        "id": 39326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006916.wav"
+            ]
+        }
+    },
+    {
+        "id": 39329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003465.wav"
+            ]
+        }
+    },
+    {
+        "id": 39331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001658.wav"
+            ]
+        }
+    },
+    {
+        "id": 39333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005295.wav"
+            ]
+        }
+    },
+    {
+        "id": 39349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich electric_bass makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002136.wav"
+            ]
+        }
+    },
+    {
+        "id": 39352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008479.wav"
+            ]
+        }
+    },
+    {
+        "id": 39353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004320.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004320.wav"
+            ]
+        }
+    },
+    {
+        "id": 39354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004614.wav"
+            ]
+        }
+    },
+    {
+        "id": 39362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001748.wav"
+            ]
+        }
+    },
+    {
+        "id": 39363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005339.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005339.wav"
+            ]
+        }
+    },
+    {
+        "id": 39366,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002482.wav"
+            ]
+        }
+    },
+    {
+        "id": 39367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004906.wav"
+            ]
+        }
+    },
+    {
+        "id": 39379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002914.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002914.wav"
+            ]
+        }
+    },
+    {
+        "id": 39381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004469.wav"
+            ]
+        }
+    },
+    {
+        "id": 39389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000122.wav"
+            ]
+        }
+    },
+    {
+        "id": 39390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000086.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000086.wav"
+            ]
+        }
+    },
+    {
+        "id": 39399,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003704.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003704.wav"
+            ]
+        }
+    },
+    {
+        "id": 39407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008922.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008922.wav"
+            ]
+        }
+    },
+    {
+        "id": 39411,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004449.wav"
+            ]
+        }
+    },
+    {
+        "id": 39416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004531.wav"
+            ]
+        }
+    },
+    {
+        "id": 39417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007109.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007109.wav"
+            ]
+        }
+    },
+    {
+        "id": 39418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006260.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006260.wav"
+            ]
+        }
+    },
+    {
+        "id": 39429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008775.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008775.wav"
+            ]
+        }
+    },
+    {
+        "id": 39434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004499.wav"
+            ]
+        }
+    },
+    {
+        "id": 39439,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007687.wav"
+            ]
+        }
+    },
+    {
+        "id": 39445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003330.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003330.wav"
+            ]
+        }
+    },
+    {
+        "id": 39449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006497.wav"
+            ]
+        }
+    },
+    {
+        "id": 39453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007711.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007711.wav"
+            ]
+        }
+    },
+    {
+        "id": 39455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007164.wav"
+            ]
+        }
+    },
+    {
+        "id": 39456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007481.wav"
+            ]
+        }
+    },
+    {
+        "id": 39466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003267.wav"
+            ]
+        }
+    },
+    {
+        "id": 39468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007472.wav"
+            ]
+        }
+    },
+    {
+        "id": 39469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006043.wav"
+            ]
+        }
+    },
+    {
+        "id": 39470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003511.wav"
+            ]
+        }
+    },
+    {
+        "id": 39474,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003061.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003061.wav"
+            ]
+        }
+    },
+    {
+        "id": 39477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001304.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001304.wav"
+            ]
+        }
+    },
+    {
+        "id": 39485,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008158.wav"
+            ]
+        }
+    },
+    {
+        "id": 39487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008743.wav"
+            ]
+        }
+    },
+    {
+        "id": 39490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004868.wav"
+            ]
+        }
+    },
+    {
+        "id": 39492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006305.wav"
+            ]
+        }
+    },
+    {
+        "id": 39493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004999.wav"
+            ]
+        }
+    },
+    {
+        "id": 39499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004301.wav"
+            ]
+        }
+    },
+    {
+        "id": 39523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005649.wav"
+            ]
+        }
+    },
+    {
+        "id": 39525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008529.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008529.wav"
+            ]
+        }
+    },
+    {
+        "id": 39532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004254.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004254.wav"
+            ]
+        }
+    },
+    {
+        "id": 39535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006151.wav"
+            ]
+        }
+    },
+    {
+        "id": 39542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007748.wav"
+            ]
+        }
+    },
+    {
+        "id": 39559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006348.wav"
+            ]
+        }
+    },
+    {
+        "id": 39561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000507.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000507.wav"
+            ]
+        }
+    },
+    {
+        "id": 39562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005373.wav"
+            ]
+        }
+    },
+    {
+        "id": 39566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005184.wav"
+            ]
+        }
+    },
+    {
+        "id": 39567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006227.wav"
+            ]
+        }
+    },
+    {
+        "id": 39615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000819.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000819.wav"
+            ]
+        }
+    },
+    {
+        "id": 39616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000613.wav"
+            ]
+        }
+    },
+    {
+        "id": 39618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005870.wav"
+            ]
+        }
+    },
+    {
+        "id": 39620,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006200.wav"
+            ]
+        }
+    },
+    {
+        "id": 39636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005722.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005722.wav"
+            ]
+        }
+    },
+    {
+        "id": 39643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003006.wav"
+            ]
+        }
+    },
+    {
+        "id": 39644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003813.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003813.wav"
+            ]
+        }
+    },
+    {
+        "id": 39655,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005827.wav"
+            ]
+        }
+    },
+    {
+        "id": 39661,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007028.wav"
+            ]
+        }
+    },
+    {
+        "id": 39665,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000136.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000136.wav"
+            ]
+        }
+    },
+    {
+        "id": 39677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006388.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006388.wav"
+            ]
+        }
+    },
+    {
+        "id": 39681,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003793.wav"
+            ]
+        }
+    },
+    {
+        "id": 39688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008420.wav"
+            ]
+        }
+    },
+    {
+        "id": 39690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008495.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008495.wav"
+            ]
+        }
+    },
+    {
+        "id": 39693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007885.wav"
+            ]
+        }
+    },
+    {
+        "id": 39699,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001994.wav"
+            ]
+        }
+    },
+    {
+        "id": 39703,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004788.wav"
+            ]
+        }
+    },
+    {
+        "id": 39707,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003334.wav"
+            ]
+        }
+    },
+    {
+        "id": 39709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008877.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008877.wav"
+            ]
+        }
+    },
+    {
+        "id": 39724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007449.wav"
+            ]
+        }
+    },
+    {
+        "id": 39730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002524.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002524.wav"
+            ]
+        }
+    },
+    {
+        "id": 39741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003347.wav"
+            ]
+        }
+    },
+    {
+        "id": 39744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004312.wav"
+            ]
+        }
+    },
+    {
+        "id": 39745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001508.wav"
+            ]
+        }
+    },
+    {
+        "id": 39753,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007333.wav"
+            ]
+        }
+    },
+    {
+        "id": 39761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006010.wav"
+            ]
+        }
+    },
+    {
+        "id": 39765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005583.wav"
+            ]
+        }
+    },
+    {
+        "id": 39766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001876.wav"
+            ]
+        }
+    },
+    {
+        "id": 39773,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006921.wav"
+            ]
+        }
+    },
+    {
+        "id": 39778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008221.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008221.wav"
+            ]
+        }
+    },
+    {
+        "id": 39779,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005565.wav"
+            ]
+        }
+    },
+    {
+        "id": 39780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005509.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005509.wav"
+            ]
+        }
+    },
+    {
+        "id": 39784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004935.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004935.wav"
+            ]
+        }
+    },
+    {
+        "id": 39788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005649.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005649.wav"
+            ]
+        }
+    },
+    {
+        "id": 39803,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003910.wav"
+            ]
+        }
+    },
+    {
+        "id": 39806,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008566.wav"
+            ]
+        }
+    },
+    {
+        "id": 39814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006393.wav"
+            ]
+        }
+    },
+    {
+        "id": 39822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006111.wav"
+            ]
+        }
+    },
+    {
+        "id": 39827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000262.wav"
+            ]
+        }
+    },
+    {
+        "id": 39842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many types of musical instruments appeared in the entire video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007693.wav"
+            ]
+        }
+    },
+    {
+        "id": 39848,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000336.wav"
+            ]
+        }
+    },
+    {
+        "id": 39857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008869.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008869.wav"
+            ]
+        }
+    },
+    {
+        "id": 39860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005037.wav"
+            ]
+        }
+    },
+    {
+        "id": 39867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments in the video did not sound from beginning to end?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006689.wav"
+            ]
+        }
+    },
+    {
+        "id": 39874,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001395.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001395.wav"
+            ]
+        }
+    },
+    {
+        "id": 39878,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many instruments are sounding in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005746.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005746.wav"
+            ]
+        }
+    },
+    {
+        "id": 39884,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006417.wav"
+            ]
+        }
+    },
+    {
+        "id": 39886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004777.wav"
+            ]
+        }
+    },
+    {
+        "id": 39890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004985.wav"
+            ]
+        }
+    },
+    {
+        "id": 39891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003891.wav"
+            ]
+        }
+    },
+    {
+        "id": 39893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001754.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001754.wav"
+            ]
+        }
+    },
+    {
+        "id": 39900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002504.wav"
+            ]
+        }
+    },
+    {
+        "id": 39902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006156.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006156.wav"
+            ]
+        }
+    },
+    {
+        "id": 39913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004559.wav"
+            ]
+        }
+    },
+    {
+        "id": 39916,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005921.wav"
+            ]
+        }
+    },
+    {
+        "id": 39926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005511.wav"
+            ]
+        }
+    },
+    {
+        "id": 39927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000538.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000538.wav"
+            ]
+        }
+    },
+    {
+        "id": 39928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002551.wav"
+            ]
+        }
+    },
+    {
+        "id": 39934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005307.wav"
+            ]
+        }
+    },
+    {
+        "id": 39937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003941.wav"
+            ]
+        }
+    },
+    {
+        "id": 39938,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007496.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007496.wav"
+            ]
+        }
+    },
+    {
+        "id": 39950,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007368.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007368.wav"
+            ]
+        }
+    },
+    {
+        "id": 39953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008101.wav"
+            ]
+        }
+    },
+    {
+        "id": 39960,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001239.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001239.wav"
+            ]
+        }
+    },
+    {
+        "id": 39966,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000861.wav"
+            ]
+        }
+    },
+    {
+        "id": 39985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004932.wav"
+            ]
+        }
+    },
+    {
+        "id": 39986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007455.wav"
+            ]
+        }
+    },
+    {
+        "id": 39989,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008768.wav"
+            ]
+        }
+    },
+    {
+        "id": 39995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007603.wav"
+            ]
+        }
+    },
+    {
+        "id": 39996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006177.wav"
+            ]
+        }
+    },
+    {
+        "id": 39999,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000089.wav"
+            ]
+        }
+    },
+    {
+        "id": 40042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002750.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002750.wav"
+            ]
+        }
+    },
+    {
+        "id": 40055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003936.wav"
+            ]
+        }
+    },
+    {
+        "id": 40059,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007462.wav"
+            ]
+        }
+    },
+    {
+        "id": 40064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000308.wav"
+            ]
+        }
+    },
+    {
+        "id": 40081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001194.wav"
+            ]
+        }
+    },
+    {
+        "id": 40090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002691.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002691.wav"
+            ]
+        }
+    },
+    {
+        "id": 40093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000919.wav"
+            ]
+        }
+    },
+    {
+        "id": 40094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001777.wav"
+            ]
+        }
+    },
+    {
+        "id": 40098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006559.wav"
+            ]
+        }
+    },
+    {
+        "id": 40099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004197.wav"
+            ]
+        }
+    },
+    {
+        "id": 40101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003070.wav"
+            ]
+        }
+    },
+    {
+        "id": 40119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003125.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003125.wav"
+            ]
+        }
+    },
+    {
+        "id": 40127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left louder than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008860.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008860.wav"
+            ]
+        }
+    },
+    {
+        "id": 40132,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005227.wav"
+            ]
+        }
+    },
+    {
+        "id": 40138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007766.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007766.wav"
+            ]
+        }
+    },
+    {
+        "id": 40140,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007325.wav"
+            ]
+        }
+    },
+    {
+        "id": 40142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002024.wav"
+            ]
+        }
+    },
+    {
+        "id": 40152,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003627.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003627.wav"
+            ]
+        }
+    },
+    {
+        "id": 40153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001477.wav"
+            ]
+        }
+    },
+    {
+        "id": 40169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000348.wav"
+            ]
+        }
+    },
+    {
+        "id": 40171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003831.wav"
+            ]
+        }
+    },
+    {
+        "id": 40173,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000487.wav"
+            ]
+        }
+    },
+    {
+        "id": 40177,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000121.wav"
+            ]
+        }
+    },
+    {
+        "id": 40179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000079.wav"
+            ]
+        }
+    },
+    {
+        "id": 40188,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002560.wav"
+            ]
+        }
+    },
+    {
+        "id": 40190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006338.wav"
+            ]
+        }
+    },
+    {
+        "id": 40192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005508.wav"
+            ]
+        }
+    },
+    {
+        "id": 40198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006930.wav"
+            ]
+        }
+    },
+    {
+        "id": 40200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000736.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000736.wav"
+            ]
+        }
+    },
+    {
+        "id": 40207,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000216.wav"
+            ]
+        }
+    },
+    {
+        "id": 40219,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001092.wav"
+            ]
+        }
+    },
+    {
+        "id": 40230,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000492.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000492.wav"
+            ]
+        }
+    },
+    {
+        "id": 40239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich guzheng makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004677.wav"
+            ]
+        }
+    },
+    {
+        "id": 40240,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000421.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000421.wav"
+            ]
+        }
+    },
+    {
+        "id": 40241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007311.wav"
+            ]
+        }
+    },
+    {
+        "id": 40249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005560.wav"
+            ]
+        }
+    },
+    {
+        "id": 40250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005972.wav"
+            ]
+        }
+    },
+    {
+        "id": 40255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005038.wav"
+            ]
+        }
+    },
+    {
+        "id": 40256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005272.wav"
+            ]
+        }
+    },
+    {
+        "id": 40257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the middle more rhythmic than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004338.wav"
+            ]
+        }
+    },
+    {
+        "id": 40260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000623.wav"
+            ]
+        }
+    },
+    {
+        "id": 40264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006493.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006493.wav"
+            ]
+        }
+    },
+    {
+        "id": 40266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003500.wav"
+            ]
+        }
+    },
+    {
+        "id": 40272,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the left louder than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000452.wav"
+            ]
+        }
+    },
+    {
+        "id": 40277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003446.wav"
+            ]
+        }
+    },
+    {
+        "id": 40288,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007950.wav"
+            ]
+        }
+    },
+    {
+        "id": 40291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000611.wav"
+            ]
+        }
+    },
+    {
+        "id": 40292,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the clarinet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007646.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007646.wav"
+            ]
+        }
+    },
+    {
+        "id": 40294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000348.wav"
+            ]
+        }
+    },
+    {
+        "id": 40313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008859.wav"
+            ]
+        }
+    },
+    {
+        "id": 40323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008701.wav"
+            ]
+        }
+    },
+    {
+        "id": 40356,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007310.wav"
+            ]
+        }
+    },
+    {
+        "id": 40360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008198.wav"
+            ]
+        }
+    },
+    {
+        "id": 40367,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006880.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006880.wav"
+            ]
+        }
+    },
+    {
+        "id": 40368,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007029.wav"
+            ]
+        }
+    },
+    {
+        "id": 40369,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004488.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004488.wav"
+            ]
+        }
+    },
+    {
+        "id": 40376,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000391.wav"
+            ]
+        }
+    },
+    {
+        "id": 40377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005552.wav"
+            ]
+        }
+    },
+    {
+        "id": 40380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008242.wav"
+            ]
+        }
+    },
+    {
+        "id": 40382,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002024.wav"
+            ]
+        }
+    },
+    {
+        "id": 40404,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele more rhythmic than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005830.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005830.wav"
+            ]
+        }
+    },
+    {
+        "id": 40405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006344.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006344.wav"
+            ]
+        }
+    },
+    {
+        "id": 40407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002772.wav"
+            ]
+        }
+    },
+    {
+        "id": 40409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005725.wav"
+            ]
+        }
+    },
+    {
+        "id": 40420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004851.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004851.wav"
+            ]
+        }
+    },
+    {
+        "id": 40433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich drum makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007643.wav"
+            ]
+        }
+    },
+    {
+        "id": 40442,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003474.wav"
+            ]
+        }
+    },
+    {
+        "id": 40446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004227.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004227.wav"
+            ]
+        }
+    },
+    {
+        "id": 40449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich banjo makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003473.wav"
+            ]
+        }
+    },
+    {
+        "id": 40450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002574.wav"
+            ]
+        }
+    },
+    {
+        "id": 40458,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008881.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008881.wav"
+            ]
+        }
+    },
+    {
+        "id": 40464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002574.wav"
+            ]
+        }
+    },
+    {
+        "id": 40466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008353.wav"
+            ]
+        }
+    },
+    {
+        "id": 40470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002508.wav"
+            ]
+        }
+    },
+    {
+        "id": 40474,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002087.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002087.wav"
+            ]
+        }
+    },
+    {
+        "id": 40495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003690.wav"
+            ]
+        }
+    },
+    {
+        "id": 40508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003854.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003854.wav"
+            ]
+        }
+    },
+    {
+        "id": 40512,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006405.wav"
+            ]
+        }
+    },
+    {
+        "id": 40528,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002955.wav"
+            ]
+        }
+    },
+    {
+        "id": 40531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000562.wav"
+            ]
+        }
+    },
+    {
+        "id": 40534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001595.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001595.wav"
+            ]
+        }
+    },
+    {
+        "id": 40535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002255.wav"
+            ]
+        }
+    },
+    {
+        "id": 40541,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002599.wav"
+            ]
+        }
+    },
+    {
+        "id": 40559,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002212.wav"
+            ]
+        }
+    },
+    {
+        "id": 40569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005720.wav"
+            ]
+        }
+    },
+    {
+        "id": 40570,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006168.wav"
+            ]
+        }
+    },
+    {
+        "id": 40575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004371.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004371.wav"
+            ]
+        }
+    },
+    {
+        "id": 40576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001432.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001432.wav"
+            ]
+        }
+    },
+    {
+        "id": 40580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000236.wav"
+            ]
+        }
+    },
+    {
+        "id": 40584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005816.wav"
+            ]
+        }
+    },
+    {
+        "id": 40598,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004663.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004663.wav"
+            ]
+        }
+    },
+    {
+        "id": 40610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000978.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000978.wav"
+            ]
+        }
+    },
+    {
+        "id": 40611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007322.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007322.wav"
+            ]
+        }
+    },
+    {
+        "id": 40614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000443.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000443.wav"
+            ]
+        }
+    },
+    {
+        "id": 40615,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004626.wav"
+            ]
+        }
+    },
+    {
+        "id": 40623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007170.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007170.wav"
+            ]
+        }
+    },
+    {
+        "id": 40630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005151.wav"
+            ]
+        }
+    },
+    {
+        "id": 40631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006405.wav"
+            ]
+        }
+    },
+    {
+        "id": 40636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006611.wav"
+            ]
+        }
+    },
+    {
+        "id": 40646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008697.wav"
+            ]
+        }
+    },
+    {
+        "id": 40647,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006940.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006940.wav"
+            ]
+        }
+    },
+    {
+        "id": 40658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004947.wav"
+            ]
+        }
+    },
+    {
+        "id": 40668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the right more rhythmic than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000079.wav"
+            ]
+        }
+    },
+    {
+        "id": 40680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005897.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005897.wav"
+            ]
+        }
+    },
+    {
+        "id": 40686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007755.wav"
+            ]
+        }
+    },
+    {
+        "id": 40693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004770.wav"
+            ]
+        }
+    },
+    {
+        "id": 40705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008146.wav"
+            ]
+        }
+    },
+    {
+        "id": 40708,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001818.wav"
+            ]
+        }
+    },
+    {
+        "id": 40713,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004073.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004073.wav"
+            ]
+        }
+    },
+    {
+        "id": 40723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007257.wav"
+            ]
+        }
+    },
+    {
+        "id": 40729,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006653.wav"
+            ]
+        }
+    },
+    {
+        "id": 40742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008092.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008092.wav"
+            ]
+        }
+    },
+    {
+        "id": 40745,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005904.wav"
+            ]
+        }
+    },
+    {
+        "id": 40756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008596.wav"
+            ]
+        }
+    },
+    {
+        "id": 40762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008903.wav"
+            ]
+        }
+    },
+    {
+        "id": 40764,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003231.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003231.wav"
+            ]
+        }
+    },
+    {
+        "id": 40766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003153.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003153.wav"
+            ]
+        }
+    },
+    {
+        "id": 40775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004447.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004447.wav"
+            ]
+        }
+    },
+    {
+        "id": 40778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001942.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001942.wav"
+            ]
+        }
+    },
+    {
+        "id": 40784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003622.wav"
+            ]
+        }
+    },
+    {
+        "id": 40785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006505.wav"
+            ]
+        }
+    },
+    {
+        "id": 40787,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002846.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002846.wav"
+            ]
+        }
+    },
+    {
+        "id": 40791,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002989.wav"
+            ]
+        }
+    },
+    {
+        "id": 40800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003609.wav"
+            ]
+        }
+    },
+    {
+        "id": 40809,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004141.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004141.wav"
+            ]
+        }
+    },
+    {
+        "id": 40810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008312.wav"
+            ]
+        }
+    },
+    {
+        "id": 40815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004122.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004122.wav"
+            ]
+        }
+    },
+    {
+        "id": 40819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002845.wav"
+            ]
+        }
+    },
+    {
+        "id": 40833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006762.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006762.wav"
+            ]
+        }
+    },
+    {
+        "id": 40837,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008735.wav"
+            ]
+        }
+    },
+    {
+        "id": 40849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005348.wav"
+            ]
+        }
+    },
+    {
+        "id": 40850,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004929.wav"
+            ]
+        }
+    },
+    {
+        "id": 40851,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002456.wav"
+            ]
+        }
+    },
+    {
+        "id": 40858,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003853.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003853.wav"
+            ]
+        }
+    },
+    {
+        "id": 40863,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006667.wav"
+            ]
+        }
+    },
+    {
+        "id": 40864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002190.wav"
+            ]
+        }
+    },
+    {
+        "id": 40865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008050.wav"
+            ]
+        }
+    },
+    {
+        "id": 40916,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002989.wav"
+            ]
+        }
+    },
+    {
+        "id": 40918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007576.wav"
+            ]
+        }
+    },
+    {
+        "id": 40924,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002753.wav"
+            ]
+        }
+    },
+    {
+        "id": 40926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006290.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006290.wav"
+            ]
+        }
+    },
+    {
+        "id": 40928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004771.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004771.wav"
+            ]
+        }
+    },
+    {
+        "id": 40936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004200.wav"
+            ]
+        }
+    },
+    {
+        "id": 40947,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002398.wav"
+            ]
+        }
+    },
+    {
+        "id": 40953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006159.wav"
+            ]
+        }
+    },
+    {
+        "id": 40955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000331.wav"
+            ]
+        }
+    },
+    {
+        "id": 40965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005023.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005023.wav"
+            ]
+        }
+    },
+    {
+        "id": 40970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005314.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005314.wav"
+            ]
+        }
+    },
+    {
+        "id": 40971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008630.wav"
+            ]
+        }
+    },
+    {
+        "id": 40973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003685.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003685.wav"
+            ]
+        }
+    },
+    {
+        "id": 40978,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005356.wav"
+            ]
+        }
+    },
+    {
+        "id": 40981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008770.wav"
+            ]
+        }
+    },
+    {
+        "id": 40983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008119.wav"
+            ]
+        }
+    },
+    {
+        "id": 40984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004594.wav"
+            ]
+        }
+    },
+    {
+        "id": 40985,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002268.wav"
+            ]
+        }
+    },
+    {
+        "id": 40987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006867.wav"
+            ]
+        }
+    },
+    {
+        "id": 40988,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004512.wav"
+            ]
+        }
+    },
+    {
+        "id": 40998,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003135.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003135.wav"
+            ]
+        }
+    },
+    {
+        "id": 41003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007318.wav"
+            ]
+        }
+    },
+    {
+        "id": 41004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006508.wav"
+            ]
+        }
+    },
+    {
+        "id": 41009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002282.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002282.wav"
+            ]
+        }
+    },
+    {
+        "id": 41010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003030.wav"
+            ]
+        }
+    },
+    {
+        "id": 41017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008245.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008245.wav"
+            ]
+        }
+    },
+    {
+        "id": 41022,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006879.wav"
+            ]
+        }
+    },
+    {
+        "id": 41027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007828.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007828.wav"
+            ]
+        }
+    },
+    {
+        "id": 41032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000351.wav"
+            ]
+        }
+    },
+    {
+        "id": 41034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008180.wav"
+            ]
+        }
+    },
+    {
+        "id": 41043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002440.wav"
+            ]
+        }
+    },
+    {
+        "id": 41045,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002207.wav"
+            ]
+        }
+    },
+    {
+        "id": 41047,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006177.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006177.wav"
+            ]
+        }
+    },
+    {
+        "id": 41048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005110.wav"
+            ]
+        }
+    },
+    {
+        "id": 41051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003919.wav"
+            ]
+        }
+    },
+    {
+        "id": 41060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004351.wav"
+            ]
+        }
+    },
+    {
+        "id": 41066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007822.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007822.wav"
+            ]
+        }
+    },
+    {
+        "id": 41080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008608.wav"
+            ]
+        }
+    },
+    {
+        "id": 41081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004531.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004531.wav"
+            ]
+        }
+    },
+    {
+        "id": 41086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005596.wav"
+            ]
+        }
+    },
+    {
+        "id": 41089,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006197.wav"
+            ]
+        }
+    },
+    {
+        "id": 41094,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007089.wav"
+            ]
+        }
+    },
+    {
+        "id": 41095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002360.wav"
+            ]
+        }
+    },
+    {
+        "id": 41096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008000.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008000.wav"
+            ]
+        }
+    },
+    {
+        "id": 41098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007696.wav"
+            ]
+        }
+    },
+    {
+        "id": 41099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007308.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007308.wav"
+            ]
+        }
+    },
+    {
+        "id": 41101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich saxophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007317.wav"
+            ]
+        }
+    },
+    {
+        "id": 41112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001045.wav"
+            ]
+        }
+    },
+    {
+        "id": 41133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007016.wav"
+            ]
+        }
+    },
+    {
+        "id": 41134,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001410.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001410.wav"
+            ]
+        }
+    },
+    {
+        "id": 41135,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006867.wav"
+            ]
+        }
+    },
+    {
+        "id": 41136,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007274.wav"
+            ]
+        }
+    },
+    {
+        "id": 41157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008980.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008980.wav"
+            ]
+        }
+    },
+    {
+        "id": 41164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008451.wav"
+            ]
+        }
+    },
+    {
+        "id": 41166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004916.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004916.wav"
+            ]
+        }
+    },
+    {
+        "id": 41167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004901.wav"
+            ]
+        }
+    },
+    {
+        "id": 41168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000406.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000406.wav"
+            ]
+        }
+    },
+    {
+        "id": 41172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006257.wav"
+            ]
+        }
+    },
+    {
+        "id": 41189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003216.wav"
+            ]
+        }
+    },
+    {
+        "id": 41192,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006476.wav"
+            ]
+        }
+    },
+    {
+        "id": 41197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005479.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005479.wav"
+            ]
+        }
+    },
+    {
+        "id": 41201,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003929.wav"
+            ]
+        }
+    },
+    {
+        "id": 41202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005787.wav"
+            ]
+        }
+    },
+    {
+        "id": 41206,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005798.wav"
+            ]
+        }
+    },
+    {
+        "id": 41208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006341.wav"
+            ]
+        }
+    },
+    {
+        "id": 41221,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008327.wav"
+            ]
+        }
+    },
+    {
+        "id": 41241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006113.wav"
+            ]
+        }
+    },
+    {
+        "id": 41251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003798.wav"
+            ]
+        }
+    },
+    {
+        "id": 41253,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006915.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006915.wav"
+            ]
+        }
+    },
+    {
+        "id": 41264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004404.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004404.wav"
+            ]
+        }
+    },
+    {
+        "id": 41268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007967.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007967.wav"
+            ]
+        }
+    },
+    {
+        "id": 41270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003298.wav"
+            ]
+        }
+    },
+    {
+        "id": 41271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008761.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008761.wav"
+            ]
+        }
+    },
+    {
+        "id": 41281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007134.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007134.wav"
+            ]
+        }
+    },
+    {
+        "id": 41290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006743.wav"
+            ]
+        }
+    },
+    {
+        "id": 41292,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001823.wav"
+            ]
+        }
+    },
+    {
+        "id": 41305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004287.wav"
+            ]
+        }
+    },
+    {
+        "id": 41308,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000229.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000229.wav"
+            ]
+        }
+    },
+    {
+        "id": 41324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001848.wav"
+            ]
+        }
+    },
+    {
+        "id": 41354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001932.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001932.wav"
+            ]
+        }
+    },
+    {
+        "id": 41355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006959.wav"
+            ]
+        }
+    },
+    {
+        "id": 41357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004249.wav"
+            ]
+        }
+    },
+    {
+        "id": 41372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002049.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002049.wav"
+            ]
+        }
+    },
+    {
+        "id": 41378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005985.wav"
+            ]
+        }
+    },
+    {
+        "id": 41379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left louder than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000298.wav"
+            ]
+        }
+    },
+    {
+        "id": 41383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007571.wav"
+            ]
+        }
+    },
+    {
+        "id": 41389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006338.wav"
+            ]
+        }
+    },
+    {
+        "id": 41390,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004917.wav"
+            ]
+        }
+    },
+    {
+        "id": 41391,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001604.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001604.wav"
+            ]
+        }
+    },
+    {
+        "id": 41396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the ukulele on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002571.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002571.wav"
+            ]
+        }
+    },
+    {
+        "id": 41397,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003295.wav"
+            ]
+        }
+    },
+    {
+        "id": 41405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005574.wav"
+            ]
+        }
+    },
+    {
+        "id": 41409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004561.wav"
+            ]
+        }
+    },
+    {
+        "id": 41410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006226.wav"
+            ]
+        }
+    },
+    {
+        "id": 41416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000455.wav"
+            ]
+        }
+    },
+    {
+        "id": 41418,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007584.wav"
+            ]
+        }
+    },
+    {
+        "id": 41420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005101.wav"
+            ]
+        }
+    },
+    {
+        "id": 41426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002732.wav"
+            ]
+        }
+    },
+    {
+        "id": 41440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002953.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002953.wav"
+            ]
+        }
+    },
+    {
+        "id": 41441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007137.wav"
+            ]
+        }
+    },
+    {
+        "id": 41445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000487.wav"
+            ]
+        }
+    },
+    {
+        "id": 41446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005446.wav"
+            ]
+        }
+    },
+    {
+        "id": 41451,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich xylophone makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000532.wav"
+            ]
+        }
+    },
+    {
+        "id": 41456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002084.wav"
+            ]
+        }
+    },
+    {
+        "id": 41464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006614.wav"
+            ]
+        }
+    },
+    {
+        "id": 41467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005970.wav"
+            ]
+        }
+    },
+    {
+        "id": 41468,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008884.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008884.wav"
+            ]
+        }
+    },
+    {
+        "id": 41477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007917.wav"
+            ]
+        }
+    },
+    {
+        "id": 41480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006295.wav"
+            ]
+        }
+    },
+    {
+        "id": 41486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005787.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005787.wav"
+            ]
+        }
+    },
+    {
+        "id": 41500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004452.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004452.wav"
+            ]
+        }
+    },
+    {
+        "id": 41503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008303.wav"
+            ]
+        }
+    },
+    {
+        "id": 41504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005729.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005729.wav"
+            ]
+        }
+    },
+    {
+        "id": 41506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008330.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008330.wav"
+            ]
+        }
+    },
+    {
+        "id": 41508,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007248.wav"
+            ]
+        }
+    },
+    {
+        "id": 41522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000081.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000081.wav"
+            ]
+        }
+    },
+    {
+        "id": 41523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005693.wav"
+            ]
+        }
+    },
+    {
+        "id": 41530,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002867.wav"
+            ]
+        }
+    },
+    {
+        "id": 41538,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004618.wav"
+            ]
+        }
+    },
+    {
+        "id": 41549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005491.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005491.wav"
+            ]
+        }
+    },
+    {
+        "id": 41555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007555.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007555.wav"
+            ]
+        }
+    },
+    {
+        "id": 41562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007670.wav"
+            ]
+        }
+    },
+    {
+        "id": 41567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002556.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002556.wav"
+            ]
+        }
+    },
+    {
+        "id": 41570,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003677.wav"
+            ]
+        }
+    },
+    {
+        "id": 41581,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003609.wav"
+            ]
+        }
+    },
+    {
+        "id": 41588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006602.wav"
+            ]
+        }
+    },
+    {
+        "id": 41602,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004939.wav"
+            ]
+        }
+    },
+    {
+        "id": 41604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008715.wav"
+            ]
+        }
+    },
+    {
+        "id": 41605,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002377.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002377.wav"
+            ]
+        }
+    },
+    {
+        "id": 41617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005383.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005383.wav"
+            ]
+        }
+    },
+    {
+        "id": 41623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005551.wav"
+            ]
+        }
+    },
+    {
+        "id": 41643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001487.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001487.wav"
+            ]
+        }
+    },
+    {
+        "id": 41644,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008101.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008101.wav"
+            ]
+        }
+    },
+    {
+        "id": 41646,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006062.wav"
+            ]
+        }
+    },
+    {
+        "id": 41648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003511.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003511.wav"
+            ]
+        }
+    },
+    {
+        "id": 41676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004614.wav"
+            ]
+        }
+    },
+    {
+        "id": 41680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004293.wav"
+            ]
+        }
+    },
+    {
+        "id": 41688,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007326.wav"
+            ]
+        }
+    },
+    {
+        "id": 41690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001806.wav"
+            ]
+        }
+    },
+    {
+        "id": 41691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007075.wav"
+            ]
+        }
+    },
+    {
+        "id": 41700,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004727.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004727.wav"
+            ]
+        }
+    },
+    {
+        "id": 41709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002622.wav"
+            ]
+        }
+    },
+    {
+        "id": 41711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005504.wav"
+            ]
+        }
+    },
+    {
+        "id": 41715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006273.wav"
+            ]
+        }
+    },
+    {
+        "id": 41717,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005175.wav"
+            ]
+        }
+    },
+    {
+        "id": 41718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003477.wav"
+            ]
+        }
+    },
+    {
+        "id": 41719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007788.wav"
+            ]
+        }
+    },
+    {
+        "id": 41720,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008700.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008700.wav"
+            ]
+        }
+    },
+    {
+        "id": 41722,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000127.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000127.wav"
+            ]
+        }
+    },
+    {
+        "id": 41724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001057.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001057.wav"
+            ]
+        }
+    },
+    {
+        "id": 41726,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007184.wav"
+            ]
+        }
+    },
+    {
+        "id": 41730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004999.wav"
+            ]
+        }
+    },
+    {
+        "id": 41735,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007208.wav"
+            ]
+        }
+    },
+    {
+        "id": 41736,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003337.wav"
+            ]
+        }
+    },
+    {
+        "id": 41750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005574.wav"
+            ]
+        }
+    },
+    {
+        "id": 41757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006959.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006959.wav"
+            ]
+        }
+    },
+    {
+        "id": 41766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000606.wav"
+            ]
+        }
+    },
+    {
+        "id": 41778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008817.wav"
+            ]
+        }
+    },
+    {
+        "id": 41780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008189.wav"
+            ]
+        }
+    },
+    {
+        "id": 41784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000566.wav"
+            ]
+        }
+    },
+    {
+        "id": 41786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000893.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000893.wav"
+            ]
+        }
+    },
+    {
+        "id": 41789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003231.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003231.wav"
+            ]
+        }
+    },
+    {
+        "id": 41790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001804.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001804.wav"
+            ]
+        }
+    },
+    {
+        "id": 41792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006619.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006619.wav"
+            ]
+        }
+    },
+    {
+        "id": 41793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008528.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008528.wav"
+            ]
+        }
+    },
+    {
+        "id": 41794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001367.wav"
+            ]
+        }
+    },
+    {
+        "id": 41797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003499.wav"
+            ]
+        }
+    },
+    {
+        "id": 41810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005345.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005345.wav"
+            ]
+        }
+    },
+    {
+        "id": 41811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002453.wav"
+            ]
+        }
+    },
+    {
+        "id": 41812,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000469.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000469.wav"
+            ]
+        }
+    },
+    {
+        "id": 41815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006602.wav"
+            ]
+        }
+    },
+    {
+        "id": 41816,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005068.wav"
+            ]
+        }
+    },
+    {
+        "id": 41817,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003894.wav"
+            ]
+        }
+    },
+    {
+        "id": 41818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002674.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002674.wav"
+            ]
+        }
+    },
+    {
+        "id": 41819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008079.wav"
+            ]
+        }
+    },
+    {
+        "id": 41823,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004200.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004200.wav"
+            ]
+        }
+    },
+    {
+        "id": 41829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000684.wav"
+            ]
+        }
+    },
+    {
+        "id": 41830,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006643.wav"
+            ]
+        }
+    },
+    {
+        "id": 41833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003533.wav"
+            ]
+        }
+    },
+    {
+        "id": 41840,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002786.wav"
+            ]
+        }
+    },
+    {
+        "id": 41841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000515.wav"
+            ]
+        }
+    },
+    {
+        "id": 41845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004030.wav"
+            ]
+        }
+    },
+    {
+        "id": 41852,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003836.wav"
+            ]
+        }
+    },
+    {
+        "id": 41882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001552.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001552.wav"
+            ]
+        }
+    },
+    {
+        "id": 41883,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004499.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004499.wav"
+            ]
+        }
+    },
+    {
+        "id": 41891,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003556.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003556.wav"
+            ]
+        }
+    },
+    {
+        "id": 41892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006721.wav"
+            ]
+        }
+    },
+    {
+        "id": 41897,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005870.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005870.wav"
+            ]
+        }
+    },
+    {
+        "id": 41901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000121.wav"
+            ]
+        }
+    },
+    {
+        "id": 41918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000256.wav"
+            ]
+        }
+    },
+    {
+        "id": 41920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006706.wav"
+            ]
+        }
+    },
+    {
+        "id": 41929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007245.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007245.wav"
+            ]
+        }
+    },
+    {
+        "id": 41930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008772.wav"
+            ]
+        }
+    },
+    {
+        "id": 41931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003597.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003597.wav"
+            ]
+        }
+    },
+    {
+        "id": 41936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005307.wav"
+            ]
+        }
+    },
+    {
+        "id": 41943,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004655.wav"
+            ]
+        }
+    },
+    {
+        "id": 41946,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004826.wav"
+            ]
+        }
+    },
+    {
+        "id": 41952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002906.wav"
+            ]
+        }
+    },
+    {
+        "id": 41953,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005942.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005942.wav"
+            ]
+        }
+    },
+    {
+        "id": 41957,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006007.wav"
+            ]
+        }
+    },
+    {
+        "id": 41961,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bassoon in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005490.wav"
+            ]
+        }
+    },
+    {
+        "id": 41963,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003614.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003614.wav"
+            ]
+        }
+    },
+    {
+        "id": 41964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005370.wav"
+            ]
+        }
+    },
+    {
+        "id": 41970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006326.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006326.wav"
+            ]
+        }
+    },
+    {
+        "id": 41973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007638.wav"
+            ]
+        }
+    },
+    {
+        "id": 41976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006539.wav"
+            ]
+        }
+    },
+    {
+        "id": 41990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007817.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007817.wav"
+            ]
+        }
+    },
+    {
+        "id": 41995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006295.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006295.wav"
+            ]
+        }
+    },
+    {
+        "id": 42002,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008764.wav"
+            ]
+        }
+    },
+    {
+        "id": 42004,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002667.wav"
+            ]
+        }
+    },
+    {
+        "id": 42008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001267.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001267.wav"
+            ]
+        }
+    },
+    {
+        "id": 42017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008050.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008050.wav"
+            ]
+        }
+    },
+    {
+        "id": 42026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding bagpipe in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008593.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008593.wav"
+            ]
+        }
+    },
+    {
+        "id": 42034,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007481.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007481.wav"
+            ]
+        }
+    },
+    {
+        "id": 42040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006472.wav"
+            ]
+        }
+    },
+    {
+        "id": 42069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007769.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007769.wav"
+            ]
+        }
+    },
+    {
+        "id": 42071,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007909.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007909.wav"
+            ]
+        }
+    },
+    {
+        "id": 42073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002386.wav"
+            ]
+        }
+    },
+    {
+        "id": 42074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008303.wav"
+            ]
+        }
+    },
+    {
+        "id": 42093,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000301.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000301.wav"
+            ]
+        }
+    },
+    {
+        "id": 42102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004255.wav"
+            ]
+        }
+    },
+    {
+        "id": 42112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding suona in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003705.wav"
+            ]
+        }
+    },
+    {
+        "id": 42113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000876.wav"
+            ]
+        }
+    },
+    {
+        "id": 42116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007184.wav"
+            ]
+        }
+    },
+    {
+        "id": 42118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007417.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007417.wav"
+            ]
+        }
+    },
+    {
+        "id": 42130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006197.wav"
+            ]
+        }
+    },
+    {
+        "id": 42133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008690.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008690.wav"
+            ]
+        }
+    },
+    {
+        "id": 42139,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005501.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005501.wav"
+            ]
+        }
+    },
+    {
+        "id": 42142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bagpipe makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002911.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002911.wav"
+            ]
+        }
+    },
+    {
+        "id": 42143,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005348.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005348.wav"
+            ]
+        }
+    },
+    {
+        "id": 42145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002906.wav"
+            ]
+        }
+    },
+    {
+        "id": 42149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004939.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004939.wav"
+            ]
+        }
+    },
+    {
+        "id": 42150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000380.wav"
+            ]
+        }
+    },
+    {
+        "id": 42164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000606.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000606.wav"
+            ]
+        }
+    },
+    {
+        "id": 42166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004370.wav"
+            ]
+        }
+    },
+    {
+        "id": 42168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006826.wav"
+            ]
+        }
+    },
+    {
+        "id": 42185,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005678.wav"
+            ]
+        }
+    },
+    {
+        "id": 42191,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005662.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005662.wav"
+            ]
+        }
+    },
+    {
+        "id": 42199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001628.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001628.wav"
+            ]
+        }
+    },
+    {
+        "id": 42200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005079.wav"
+            ]
+        }
+    },
+    {
+        "id": 42204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005184.wav"
+            ]
+        }
+    },
+    {
+        "id": 42209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003278.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003278.wav"
+            ]
+        }
+    },
+    {
+        "id": 42210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008905.wav"
+            ]
+        }
+    },
+    {
+        "id": 42224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003992.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003992.wav"
+            ]
+        }
+    },
+    {
+        "id": 42234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000518.wav"
+            ]
+        }
+    },
+    {
+        "id": 42236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004366.wav"
+            ]
+        }
+    },
+    {
+        "id": 42239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004341.wav"
+            ]
+        }
+    },
+    {
+        "id": 42241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006734.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006734.wav"
+            ]
+        }
+    },
+    {
+        "id": 42246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004242.wav"
+            ]
+        }
+    },
+    {
+        "id": 42263,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008327.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008327.wav"
+            ]
+        }
+    },
+    {
+        "id": 42265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005720.wav"
+            ]
+        }
+    },
+    {
+        "id": 42266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007233.wav"
+            ]
+        }
+    },
+    {
+        "id": 42271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000584.wav"
+            ]
+        }
+    },
+    {
+        "id": 42284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004019.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004019.wav"
+            ]
+        }
+    },
+    {
+        "id": 42286,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007976.wav"
+            ]
+        }
+    },
+    {
+        "id": 42312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005413.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005413.wav"
+            ]
+        }
+    },
+    {
+        "id": 42320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007075.wav"
+            ]
+        }
+    },
+    {
+        "id": 42326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008263.wav"
+            ]
+        }
+    },
+    {
+        "id": 42333,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003350.wav"
+            ]
+        }
+    },
+    {
+        "id": 42343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007848.wav"
+            ]
+        }
+    },
+    {
+        "id": 42346,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding guzheng in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006832.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006832.wav"
+            ]
+        }
+    },
+    {
+        "id": 42354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007638.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007638.wav"
+            ]
+        }
+    },
+    {
+        "id": 42357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004962.wav"
+            ]
+        }
+    },
+    {
+        "id": 42379,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the left louder than the bagpipe on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002713.wav"
+            ]
+        }
+    },
+    {
+        "id": 42380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000741.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000741.wav"
+            ]
+        }
+    },
+    {
+        "id": 42402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008639.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008639.wav"
+            ]
+        }
+    },
+    {
+        "id": 42409,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005385.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005385.wav"
+            ]
+        }
+    },
+    {
+        "id": 42415,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001324.wav"
+            ]
+        }
+    },
+    {
+        "id": 42423,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008312.wav"
+            ]
+        }
+    },
+    {
+        "id": 42433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003816.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003816.wav"
+            ]
+        }
+    },
+    {
+        "id": 42435,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006772.wav"
+            ]
+        }
+    },
+    {
+        "id": 42452,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000978.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000978.wav"
+            ]
+        }
+    },
+    {
+        "id": 42454,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007952.wav"
+            ]
+        }
+    },
+    {
+        "id": 42457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003601.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003601.wav"
+            ]
+        }
+    },
+    {
+        "id": 42461,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004608.wav"
+            ]
+        }
+    },
+    {
+        "id": 42463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002534.wav"
+            ]
+        }
+    },
+    {
+        "id": 42464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich bassoon makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005490.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005490.wav"
+            ]
+        }
+    },
+    {
+        "id": 42474,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004473.wav"
+            ]
+        }
+    },
+    {
+        "id": 42479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding saxophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004896.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004896.wav"
+            ]
+        }
+    },
+    {
+        "id": 42484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006340.wav"
+            ]
+        }
+    },
+    {
+        "id": 42487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the left louder than the suona on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003075.wav"
+            ]
+        }
+    },
+    {
+        "id": 42491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008347.wav"
+            ]
+        }
+    },
+    {
+        "id": 42495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004212.wav"
+            ]
+        }
+    },
+    {
+        "id": 42503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005849.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005849.wav"
+            ]
+        }
+    },
+    {
+        "id": 42525,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005985.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005985.wav"
+            ]
+        }
+    },
+    {
+        "id": 42528,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004024.wav"
+            ]
+        }
+    },
+    {
+        "id": 42545,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002556.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002556.wav"
+            ]
+        }
+    },
+    {
+        "id": 42547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008743.wav"
+            ]
+        }
+    },
+    {
+        "id": 42552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000664.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000664.wav"
+            ]
+        }
+    },
+    {
+        "id": 42556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005273.wav"
+            ]
+        }
+    },
+    {
+        "id": 42560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008466.wav"
+            ]
+        }
+    },
+    {
+        "id": 42564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005843.wav"
+            ]
+        }
+    },
+    {
+        "id": 42565,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005393.wav"
+            ]
+        }
+    },
+    {
+        "id": 42575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006533.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006533.wav"
+            ]
+        }
+    },
+    {
+        "id": 42583,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004948.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004948.wav"
+            ]
+        }
+    },
+    {
+        "id": 42591,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005750.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005750.wav"
+            ]
+        }
+    },
+    {
+        "id": 42596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000365.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000365.wav"
+            ]
+        }
+    },
+    {
+        "id": 42600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002717.wav"
+            ]
+        }
+    },
+    {
+        "id": 42653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004255.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004255.wav"
+            ]
+        }
+    },
+    {
+        "id": 42662,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007421.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007421.wav"
+            ]
+        }
+    },
+    {
+        "id": 42664,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008375.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008375.wav"
+            ]
+        }
+    },
+    {
+        "id": 42667,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002927.wav"
+            ]
+        }
+    },
+    {
+        "id": 42668,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005574.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005574.wav"
+            ]
+        }
+    },
+    {
+        "id": 42672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008374.wav"
+            ]
+        }
+    },
+    {
+        "id": 42677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005575.wav"
+            ]
+        }
+    },
+    {
+        "id": 42680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003268.wav"
+            ]
+        }
+    },
+    {
+        "id": 42701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000876.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000876.wav"
+            ]
+        }
+    },
+    {
+        "id": 42704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding xylophone in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008400.wav"
+            ]
+        }
+    },
+    {
+        "id": 42716,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007139.wav"
+            ]
+        }
+    },
+    {
+        "id": 42723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006268.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006268.wav"
+            ]
+        }
+    },
+    {
+        "id": 42730,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005889.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005889.wav"
+            ]
+        }
+    },
+    {
+        "id": 42734,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000119.wav"
+            ]
+        }
+    },
+    {
+        "id": 42743,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004901.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004901.wav"
+            ]
+        }
+    },
+    {
+        "id": 42757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004351.wav"
+            ]
+        }
+    },
+    {
+        "id": 42765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005220.wav"
+            ]
+        }
+    },
+    {
+        "id": 42774,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006318.wav"
+            ]
+        }
+    },
+    {
+        "id": 42776,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004814.wav"
+            ]
+        }
+    },
+    {
+        "id": 42785,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007695.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007695.wav"
+            ]
+        }
+    },
+    {
+        "id": 42789,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005660.wav"
+            ]
+        }
+    },
+    {
+        "id": 42796,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001214.wav"
+            ]
+        }
+    },
+    {
+        "id": 42802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003236.wav"
+            ]
+        }
+    },
+    {
+        "id": 42825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007752.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007752.wav"
+            ]
+        }
+    },
+    {
+        "id": 42831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008229.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008229.wav"
+            ]
+        }
+    },
+    {
+        "id": 42834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006721.wav"
+            ]
+        }
+    },
+    {
+        "id": 42841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000453.wav"
+            ]
+        }
+    },
+    {
+        "id": 42845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003569.wav"
+            ]
+        }
+    },
+    {
+        "id": 42849,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding flute in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008279.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008279.wav"
+            ]
+        }
+    },
+    {
+        "id": 42854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007965.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007965.wav"
+            ]
+        }
+    },
+    {
+        "id": 42864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004384.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004384.wav"
+            ]
+        }
+    },
+    {
+        "id": 42866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005931.wav"
+            ]
+        }
+    },
+    {
+        "id": 42871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004374.wav"
+            ]
+        }
+    },
+    {
+        "id": 42885,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006159.wav"
+            ]
+        }
+    },
+    {
+        "id": 42886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007158.wav"
+            ]
+        }
+    },
+    {
+        "id": 42893,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding cello in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004530.wav"
+            ]
+        }
+    },
+    {
+        "id": 42896,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008608.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008608.wav"
+            ]
+        }
+    },
+    {
+        "id": 42898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008097.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008097.wav"
+            ]
+        }
+    },
+    {
+        "id": 42908,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001879.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001879.wav"
+            ]
+        }
+    },
+    {
+        "id": 42910,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding violin in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008513.wav"
+            ]
+        }
+    },
+    {
+        "id": 42917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003207.wav"
+            ]
+        }
+    },
+    {
+        "id": 42928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008030.wav"
+            ]
+        }
+    },
+    {
+        "id": 42933,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding clarinet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005683.wav"
+            ]
+        }
+    },
+    {
+        "id": 42942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005805.wav"
+            ]
+        }
+    },
+    {
+        "id": 42948,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002258.wav"
+            ]
+        }
+    },
+    {
+        "id": 42951,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004717.wav"
+            ]
+        }
+    },
+    {
+        "id": 42958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007848.wav"
+            ]
+        }
+    },
+    {
+        "id": 42959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004997.wav"
+            ]
+        }
+    },
+    {
+        "id": 42965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001611.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001611.wav"
+            ]
+        }
+    },
+    {
+        "id": 42969,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004626.wav"
+            ]
+        }
+    },
+    {
+        "id": 42971,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003584.wav"
+            ]
+        }
+    },
+    {
+        "id": 42975,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding drum in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008219.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008219.wav"
+            ]
+        }
+    },
+    {
+        "id": 42977,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005466.wav"
+            ]
+        }
+    },
+    {
+        "id": 42982,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002956.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002956.wav"
+            ]
+        }
+    },
+    {
+        "id": 42986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003902.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003902.wav"
+            ]
+        }
+    },
+    {
+        "id": 42987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000989.wav"
+            ]
+        }
+    },
+    {
+        "id": 42990,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007755.wav"
+            ]
+        }
+    },
+    {
+        "id": 42991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008904.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008904.wav"
+            ]
+        }
+    },
+    {
+        "id": 43011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007026.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007026.wav"
+            ]
+        }
+    },
+    {
+        "id": 43019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006456.wav"
+            ]
+        }
+    },
+    {
+        "id": 43030,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002929.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002929.wav"
+            ]
+        }
+    },
+    {
+        "id": 43032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006667.wav"
+            ]
+        }
+    },
+    {
+        "id": 43037,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000544.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000544.wav"
+            ]
+        }
+    },
+    {
+        "id": 43042,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005225.wav"
+            ]
+        }
+    },
+    {
+        "id": 43048,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding accordion in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004702.wav"
+            ]
+        }
+    },
+    {
+        "id": 43049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002717.wav"
+            ]
+        }
+    },
+    {
+        "id": 43057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000431.wav"
+            ]
+        }
+    },
+    {
+        "id": 43063,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006006.wav"
+            ]
+        }
+    },
+    {
+        "id": 43075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding tuba in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003273.wav"
+            ]
+        }
+    },
+    {
+        "id": 43096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008588.wav"
+            ]
+        }
+    },
+    {
+        "id": 43101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008868.wav"
+            ]
+        }
+    },
+    {
+        "id": 43102,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the middle of piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008497.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008497.wav"
+            ]
+        }
+    },
+    {
+        "id": 43103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003464.wav"
+            ]
+        }
+    },
+    {
+        "id": 43109,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding ukulele in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003369.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003369.wav"
+            ]
+        }
+    },
+    {
+        "id": 43112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007670.wav"
+            ]
+        }
+    },
+    {
+        "id": 43129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding congas in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005995.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005995.wav"
+            ]
+        }
+    },
+    {
+        "id": 43130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007842.wav"
+            ]
+        }
+    },
+    {
+        "id": 43134,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005220.wav"
+            ]
+        }
+    },
+    {
+        "id": 43145,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008755.wav"
+            ]
+        }
+    },
+    {
+        "id": 43148,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005894.wav"
+            ]
+        }
+    },
+    {
+        "id": 43153,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003760.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003760.wav"
+            ]
+        }
+    },
+    {
+        "id": 43156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001214.wav"
+            ]
+        }
+    },
+    {
+        "id": 43162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003378.wav"
+            ]
+        }
+    },
+    {
+        "id": 43164,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding pipa in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004041.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004041.wav"
+            ]
+        }
+    },
+    {
+        "id": 43166,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007584.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007584.wav"
+            ]
+        }
+    },
+    {
+        "id": 43168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006340.wav"
+            ]
+        }
+    },
+    {
+        "id": 43171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding trumpet in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005211.wav"
+            ]
+        }
+    },
+    {
+        "id": 43174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right louder than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002622.wav"
+            ]
+        }
+    },
+    {
+        "id": 43178,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006311.wav"
+            ]
+        }
+    },
+    {
+        "id": 43183,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the right of tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002826.wav"
+            ]
+        }
+    },
+    {
+        "id": 43190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007788.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007788.wav"
+            ]
+        }
+    },
+    {
+        "id": 43193,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007412.wav"
+            ]
+        }
+    },
+    {
+        "id": 43198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003256.wav"
+            ]
+        }
+    },
+    {
+        "id": 43218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006262.wav"
+            ]
+        }
+    },
+    {
+        "id": 43239,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding erhu in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004697.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004697.wav"
+            ]
+        }
+    },
+    {
+        "id": 43269,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008321.wav"
+            ]
+        }
+    },
+    {
+        "id": 43271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006721.wav"
+            ]
+        }
+    },
+    {
+        "id": 43274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004111.wav"
+            ]
+        }
+    },
+    {
+        "id": 43275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding banjo in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008084.wav"
+            ]
+        }
+    },
+    {
+        "id": 43277,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding piano in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007397.wav"
+            ]
+        }
+    },
+    {
+        "id": 43282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002448.wav"
+            ]
+        }
+    },
+    {
+        "id": 43283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005504.wav"
+            ]
+        }
+    },
+    {
+        "id": 43284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001575.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001575.wav"
+            ]
+        }
+    },
+    {
+        "id": 43289,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005401.wav"
+            ]
+        }
+    },
+    {
+        "id": 43290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004111.wav"
+            ]
+        }
+    },
+    {
+        "id": 43310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008835.wav"
+            ]
+        }
+    },
+    {
+        "id": 43320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the instrument on the left of drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004805.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004805.wav"
+            ]
+        }
+    },
+    {
+        "id": 43329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat kind of musical instrument is it?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006083.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006083.wav"
+            ]
+        }
+    },
+    {
+        "id": 43330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005931.wav"
+            ]
+        }
+    },
+    {
+        "id": 43334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002386.wav"
+            ]
+        }
+    },
+    {
+        "id": 43339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nHow many sounding acoustic_guitar in the video\uff1f\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004293.wav"
+            ]
+        }
+    },
+    {
+        "id": 43354,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006046.wav"
+            ]
+        }
+    },
+    {
+        "id": 43365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001367.wav"
+            ]
+        }
+    },
+    {
+        "id": 43370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007705.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007705.wav"
+            ]
+        }
+    },
+    {
+        "id": 43384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003075.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003075.wav"
+            ]
+        }
+    },
+    {
+        "id": 43385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008431.wav"
+            ]
+        }
+    },
+    {
+        "id": 43410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001168.wav"
+            ]
+        }
+    },
+    {
+        "id": 43417,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006456.wav"
+            ]
+        }
+    },
+    {
+        "id": 43425,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007594.wav"
+            ]
+        }
+    },
+    {
+        "id": 43428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004024.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004024.wav"
+            ]
+        }
+    },
+    {
+        "id": 43438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005737.wav"
+            ]
+        }
+    },
+    {
+        "id": 43448,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000124.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000124.wav"
+            ]
+        }
+    },
+    {
+        "id": 43469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005226.wav"
+            ]
+        }
+    },
+    {
+        "id": 43482,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006706.wav"
+            ]
+        }
+    },
+    {
+        "id": 43486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006172.wav"
+            ]
+        }
+    },
+    {
+        "id": 43488,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001609.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001609.wav"
+            ]
+        }
+    },
+    {
+        "id": 43491,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003173.wav"
+            ]
+        }
+    },
+    {
+        "id": 43493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008513.wav"
+            ]
+        }
+    },
+    {
+        "id": 43499,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008257.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008257.wav"
+            ]
+        }
+    },
+    {
+        "id": 43502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006908.wav"
+            ]
+        }
+    },
+    {
+        "id": 43503,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004665.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004665.wav"
+            ]
+        }
+    },
+    {
+        "id": 43513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004989.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004989.wav"
+            ]
+        }
+    },
+    {
+        "id": 43520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005154.wav"
+            ]
+        }
+    },
+    {
+        "id": 43528,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003674.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003674.wav"
+            ]
+        }
+    },
+    {
+        "id": 43529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002451.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002451.wav"
+            ]
+        }
+    },
+    {
+        "id": 43532,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008321.wav"
+            ]
+        }
+    },
+    {
+        "id": 43539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005709.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005709.wav"
+            ]
+        }
+    },
+    {
+        "id": 43547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005894.wav"
+            ]
+        }
+    },
+    {
+        "id": 43565,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003334.wav"
+            ]
+        }
+    },
+    {
+        "id": 43568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005111.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005111.wav"
+            ]
+        }
+    },
+    {
+        "id": 43569,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich violin makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005225.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005225.wav"
+            ]
+        }
+    },
+    {
+        "id": 43577,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005236.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005236.wav"
+            ]
+        }
+    },
+    {
+        "id": 43614,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005933.wav"
+            ]
+        }
+    },
+    {
+        "id": 43624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005894.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005894.wav"
+            ]
+        }
+    },
+    {
+        "id": 43637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008653.wav"
+            ]
+        }
+    },
+    {
+        "id": 43639,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008350.wav"
+            ]
+        }
+    },
+    {
+        "id": 43641,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007412.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007412.wav"
+            ]
+        }
+    },
+    {
+        "id": 43650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005994.wav"
+            ]
+        }
+    },
+    {
+        "id": 43651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008072.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008072.wav"
+            ]
+        }
+    },
+    {
+        "id": 43680,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002247.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002247.wav"
+            ]
+        }
+    },
+    {
+        "id": 43697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007161.wav"
+            ]
+        }
+    },
+    {
+        "id": 43702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000500.wav"
+            ]
+        }
+    },
+    {
+        "id": 43704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001195.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001195.wav"
+            ]
+        }
+    },
+    {
+        "id": 43719,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000470.wav"
+            ]
+        }
+    },
+    {
+        "id": 43742,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000867.wav"
+            ]
+        }
+    },
+    {
+        "id": 43747,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001202.wav"
+            ]
+        }
+    },
+    {
+        "id": 43762,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001305.wav"
+            ]
+        }
+    },
+    {
+        "id": 43766,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001188.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001188.wav"
+            ]
+        }
+    },
+    {
+        "id": 43778,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000390.wav"
+            ]
+        }
+    },
+    {
+        "id": 43786,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001239.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001239.wav"
+            ]
+        }
+    },
+    {
+        "id": 43788,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001362.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001362.wav"
+            ]
+        }
+    },
+    {
+        "id": 43810,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000873.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000873.wav"
+            ]
+        }
+    },
+    {
+        "id": 43811,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001117.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001117.wav"
+            ]
+        }
+    },
+    {
+        "id": 43813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000812.wav"
+            ]
+        }
+    },
+    {
+        "id": 43818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000405.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000405.wav"
+            ]
+        }
+    },
+    {
+        "id": 43822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000966.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000966.wav"
+            ]
+        }
+    },
+    {
+        "id": 43825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000630.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000630.wav"
+            ]
+        }
+    },
+    {
+        "id": 43826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000443.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000443.wav"
+            ]
+        }
+    },
+    {
+        "id": 43828,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000737.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000737.wav"
+            ]
+        }
+    },
+    {
+        "id": 43829,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001243.wav"
+            ]
+        }
+    },
+    {
+        "id": 43844,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001270.wav"
+            ]
+        }
+    },
+    {
+        "id": 43854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000979.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000979.wav"
+            ]
+        }
+    },
+    {
+        "id": 43856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001212.wav"
+            ]
+        }
+    },
+    {
+        "id": 43857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000521.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000521.wav"
+            ]
+        }
+    },
+    {
+        "id": 43859,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000307.wav"
+            ]
+        }
+    },
+    {
+        "id": 43877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001298.wav"
+            ]
+        }
+    },
+    {
+        "id": 43890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001456.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001456.wav"
+            ]
+        }
+    },
+    {
+        "id": 43895,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000895.wav"
+            ]
+        }
+    },
+    {
+        "id": 43917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001137.wav"
+            ]
+        }
+    },
+    {
+        "id": 43926,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000018.wav"
+            ]
+        }
+    },
+    {
+        "id": 43927,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the bagpipe?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000472.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000472.wav"
+            ]
+        }
+    },
+    {
+        "id": 43936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001000.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001000.wav"
+            ]
+        }
+    },
+    {
+        "id": 43942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000009.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000009.wav"
+            ]
+        }
+    },
+    {
+        "id": 43952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the middle more rhythmic than the drum on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001292.wav"
+            ]
+        }
+    },
+    {
+        "id": 43959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000895.wav"
+            ]
+        }
+    },
+    {
+        "id": 43973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000899.wav"
+            ]
+        }
+    },
+    {
+        "id": 43976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001942.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001942.wav"
+            ]
+        }
+    },
+    {
+        "id": 43979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000446.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000446.wav"
+            ]
+        }
+    },
+    {
+        "id": 43994,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000565.wav"
+            ]
+        }
+    },
+    {
+        "id": 44010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000414.wav"
+            ]
+        }
+    },
+    {
+        "id": 44011,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000906.wav"
+            ]
+        }
+    },
+    {
+        "id": 44021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001407.wav"
+            ]
+        }
+    },
+    {
+        "id": 44024,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000818.wav"
+            ]
+        }
+    },
+    {
+        "id": 44032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000014.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000014.wav"
+            ]
+        }
+    },
+    {
+        "id": 44037,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000990.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000990.wav"
+            ]
+        }
+    },
+    {
+        "id": 44054,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000265.wav"
+            ]
+        }
+    },
+    {
+        "id": 44060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001466.wav"
+            ]
+        }
+    },
+    {
+        "id": 44064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the last sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001172.wav"
+            ]
+        }
+    },
+    {
+        "id": 44072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000416.wav"
+            ]
+        }
+    },
+    {
+        "id": 44075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000143.wav"
+            ]
+        }
+    },
+    {
+        "id": 44076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001666.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001666.wav"
+            ]
+        }
+    },
+    {
+        "id": 44082,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003827.wav"
+            ]
+        }
+    },
+    {
+        "id": 44083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001012.wav"
+            ]
+        }
+    },
+    {
+        "id": 44086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006130.wav"
+            ]
+        }
+    },
+    {
+        "id": 44087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000626.wav"
+            ]
+        }
+    },
+    {
+        "id": 44097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001261.wav"
+            ]
+        }
+    },
+    {
+        "id": 44103,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000913.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000913.wav"
+            ]
+        }
+    },
+    {
+        "id": 44106,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000631.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000631.wav"
+            ]
+        }
+    },
+    {
+        "id": 44107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000500.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000500.wav"
+            ]
+        }
+    },
+    {
+        "id": 44112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001083.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001083.wav"
+            ]
+        }
+    },
+    {
+        "id": 44117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000321.wav"
+            ]
+        }
+    },
+    {
+        "id": 44118,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008176.wav"
+            ]
+        }
+    },
+    {
+        "id": 44124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001037.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001037.wav"
+            ]
+        }
+    },
+    {
+        "id": 44127,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004865.wav"
+            ]
+        }
+    },
+    {
+        "id": 44142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000865.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000865.wav"
+            ]
+        }
+    },
+    {
+        "id": 44150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001113.wav"
+            ]
+        }
+    },
+    {
+        "id": 44157,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000521.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000521.wav"
+            ]
+        }
+    },
+    {
+        "id": 44161,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006733.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006733.wav"
+            ]
+        }
+    },
+    {
+        "id": 44162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003484.wav"
+            ]
+        }
+    },
+    {
+        "id": 44169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000807.wav"
+            ]
+        }
+    },
+    {
+        "id": 44171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000703.wav"
+            ]
+        }
+    },
+    {
+        "id": 44172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000731.wav"
+            ]
+        }
+    },
+    {
+        "id": 44174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001164.wav"
+            ]
+        }
+    },
+    {
+        "id": 44182,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001263.wav"
+            ]
+        }
+    },
+    {
+        "id": 44198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000226.wav"
+            ]
+        }
+    },
+    {
+        "id": 44200,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000895.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000895.wav"
+            ]
+        }
+    },
+    {
+        "id": 44201,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000823.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000823.wav"
+            ]
+        }
+    },
+    {
+        "id": 44209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001182.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001182.wav"
+            ]
+        }
+    },
+    {
+        "id": 44213,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000899.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000899.wav"
+            ]
+        }
+    },
+    {
+        "id": 44218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000312.wav"
+            ]
+        }
+    },
+    {
+        "id": 44222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000324.wav"
+            ]
+        }
+    },
+    {
+        "id": 44223,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000919.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000919.wav"
+            ]
+        }
+    },
+    {
+        "id": 44226,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000727.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000727.wav"
+            ]
+        }
+    },
+    {
+        "id": 44235,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000235.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000235.wav"
+            ]
+        }
+    },
+    {
+        "id": 44248,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000319.wav"
+            ]
+        }
+    },
+    {
+        "id": 44249,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001343.wav"
+            ]
+        }
+    },
+    {
+        "id": 44250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005615.wav"
+            ]
+        }
+    },
+    {
+        "id": 44258,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007228.wav"
+            ]
+        }
+    },
+    {
+        "id": 44264,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001125.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001125.wav"
+            ]
+        }
+    },
+    {
+        "id": 44265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000194.wav"
+            ]
+        }
+    },
+    {
+        "id": 44271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007123.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007123.wav"
+            ]
+        }
+    },
+    {
+        "id": 44275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001629.wav"
+            ]
+        }
+    },
+    {
+        "id": 44288,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001025.wav"
+            ]
+        }
+    },
+    {
+        "id": 44306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000240.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000240.wav"
+            ]
+        }
+    },
+    {
+        "id": 44309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000207.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000207.wav"
+            ]
+        }
+    },
+    {
+        "id": 44310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000807.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000807.wav"
+            ]
+        }
+    },
+    {
+        "id": 44318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000398.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000398.wav"
+            ]
+        }
+    },
+    {
+        "id": 44329,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000826.wav"
+            ]
+        }
+    },
+    {
+        "id": 44331,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001071.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001071.wav"
+            ]
+        }
+    },
+    {
+        "id": 44338,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001433.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001433.wav"
+            ]
+        }
+    },
+    {
+        "id": 44344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the middle louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001629.wav"
+            ]
+        }
+    },
+    {
+        "id": 44347,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000280.wav"
+            ]
+        }
+    },
+    {
+        "id": 44357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001158.wav"
+            ]
+        }
+    },
+    {
+        "id": 44370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001234.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001234.wav"
+            ]
+        }
+    },
+    {
+        "id": 44373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000566.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000566.wav"
+            ]
+        }
+    },
+    {
+        "id": 44374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000543.wav"
+            ]
+        }
+    },
+    {
+        "id": 44375,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008373.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008373.wav"
+            ]
+        }
+    },
+    {
+        "id": 44380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000415.wav"
+            ]
+        }
+    },
+    {
+        "id": 44385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006261.wav"
+            ]
+        }
+    },
+    {
+        "id": 44387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001414.wav"
+            ]
+        }
+    },
+    {
+        "id": 44389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000211.wav"
+            ]
+        }
+    },
+    {
+        "id": 44392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000217.wav"
+            ]
+        }
+    },
+    {
+        "id": 44398,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006680.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006680.wav"
+            ]
+        }
+    },
+    {
+        "id": 44400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000142.wav"
+            ]
+        }
+    },
+    {
+        "id": 44401,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003504.wav"
+            ]
+        }
+    },
+    {
+        "id": 44403,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001148.wav"
+            ]
+        }
+    },
+    {
+        "id": 44406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005716.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005716.wav"
+            ]
+        }
+    },
+    {
+        "id": 44410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008328.wav"
+            ]
+        }
+    },
+    {
+        "id": 44420,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001347.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001347.wav"
+            ]
+        }
+    },
+    {
+        "id": 44432,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000696.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000696.wav"
+            ]
+        }
+    },
+    {
+        "id": 44434,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000885.wav"
+            ]
+        }
+    },
+    {
+        "id": 44444,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001352.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001352.wav"
+            ]
+        }
+    },
+    {
+        "id": 44449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000958.wav"
+            ]
+        }
+    },
+    {
+        "id": 44451,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left louder than the electric_bass on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000923.wav"
+            ]
+        }
+    },
+    {
+        "id": 44462,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000622.wav"
+            ]
+        }
+    },
+    {
+        "id": 44476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000605.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000605.wav"
+            ]
+        }
+    },
+    {
+        "id": 44483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001374.wav"
+            ]
+        }
+    },
+    {
+        "id": 44490,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001468.wav"
+            ]
+        }
+    },
+    {
+        "id": 44495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001212.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001212.wav"
+            ]
+        }
+    },
+    {
+        "id": 44496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000842.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000842.wav"
+            ]
+        }
+    },
+    {
+        "id": 44504,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005755.wav"
+            ]
+        }
+    },
+    {
+        "id": 44509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001374.wav"
+            ]
+        }
+    },
+    {
+        "id": 44511,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000930.wav"
+            ]
+        }
+    },
+    {
+        "id": 44515,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004414.wav"
+            ]
+        }
+    },
+    {
+        "id": 44518,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003317.wav"
+            ]
+        }
+    },
+    {
+        "id": 44519,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001293.wav"
+            ]
+        }
+    },
+    {
+        "id": 44521,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001154.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001154.wav"
+            ]
+        }
+    },
+    {
+        "id": 44523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000588.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000588.wav"
+            ]
+        }
+    },
+    {
+        "id": 44553,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000250.wav"
+            ]
+        }
+    },
+    {
+        "id": 44556,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000759.wav"
+            ]
+        }
+    },
+    {
+        "id": 44568,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001413.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001413.wav"
+            ]
+        }
+    },
+    {
+        "id": 44570,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000641.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000641.wav"
+            ]
+        }
+    },
+    {
+        "id": 44573,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007530.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007530.wav"
+            ]
+        }
+    },
+    {
+        "id": 44575,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000975.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000975.wav"
+            ]
+        }
+    },
+    {
+        "id": 44580,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000485.wav"
+            ]
+        }
+    },
+    {
+        "id": 44596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001181.wav"
+            ]
+        }
+    },
+    {
+        "id": 44599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the bagpipe on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001228.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001228.wav"
+            ]
+        }
+    },
+    {
+        "id": 44600,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000393.wav"
+            ]
+        }
+    },
+    {
+        "id": 44604,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001440.wav"
+            ]
+        }
+    },
+    {
+        "id": 44618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008561.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008561.wav"
+            ]
+        }
+    },
+    {
+        "id": 44634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001025.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001025.wav"
+            ]
+        }
+    },
+    {
+        "id": 44636,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005780.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005780.wav"
+            ]
+        }
+    },
+    {
+        "id": 44651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001317.wav"
+            ]
+        }
+    },
+    {
+        "id": 44658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000923.wav"
+            ]
+        }
+    },
+    {
+        "id": 44666,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001291.wav"
+            ]
+        }
+    },
+    {
+        "id": 44669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000659.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000659.wav"
+            ]
+        }
+    },
+    {
+        "id": 44671,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000172.wav"
+            ]
+        }
+    },
+    {
+        "id": 44672,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000356.wav"
+            ]
+        }
+    },
+    {
+        "id": 44677,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001425.wav"
+            ]
+        }
+    },
+    {
+        "id": 44679,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001636.wav"
+            ]
+        }
+    },
+    {
+        "id": 44693,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006146.wav"
+            ]
+        }
+    },
+    {
+        "id": 44702,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001068.wav"
+            ]
+        }
+    },
+    {
+        "id": 44709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001293.wav"
+            ]
+        }
+    },
+    {
+        "id": 44711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000767.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000767.wav"
+            ]
+        }
+    },
+    {
+        "id": 44718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000291.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000291.wav"
+            ]
+        }
+    },
+    {
+        "id": 44722,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000549.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000549.wav"
+            ]
+        }
+    },
+    {
+        "id": 44723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000820.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000820.wav"
+            ]
+        }
+    },
+    {
+        "id": 44739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000263.wav"
+            ]
+        }
+    },
+    {
+        "id": 44744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004798.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004798.wav"
+            ]
+        }
+    },
+    {
+        "id": 44770,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000300.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000300.wav"
+            ]
+        }
+    },
+    {
+        "id": 44802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000949.wav"
+            ]
+        }
+    },
+    {
+        "id": 44808,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000422.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000422.wav"
+            ]
+        }
+    },
+    {
+        "id": 44814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001382.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001382.wav"
+            ]
+        }
+    },
+    {
+        "id": 44818,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000666.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000666.wav"
+            ]
+        }
+    },
+    {
+        "id": 44833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001306.wav"
+            ]
+        }
+    },
+    {
+        "id": 44834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001210.wav"
+            ]
+        }
+    },
+    {
+        "id": 44846,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000461.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000461.wav"
+            ]
+        }
+    },
+    {
+        "id": 44854,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000732.wav"
+            ]
+        }
+    },
+    {
+        "id": 44856,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone louder than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000557.wav"
+            ]
+        }
+    },
+    {
+        "id": 44865,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000683.wav"
+            ]
+        }
+    },
+    {
+        "id": 44880,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000618.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000618.wav"
+            ]
+        }
+    },
+    {
+        "id": 44887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000699.wav"
+            ]
+        }
+    },
+    {
+        "id": 44894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000167.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000167.wav"
+            ]
+        }
+    },
+    {
+        "id": 44900,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001341.wav"
+            ]
+        }
+    },
+    {
+        "id": 44901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000318.wav"
+            ]
+        }
+    },
+    {
+        "id": 44903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001356.wav"
+            ]
+        }
+    },
+    {
+        "id": 44913,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001454.wav"
+            ]
+        }
+    },
+    {
+        "id": 44925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000900.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000900.wav"
+            ]
+        }
+    },
+    {
+        "id": 44928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002039.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002039.wav"
+            ]
+        }
+    },
+    {
+        "id": 44939,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000645.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000645.wav"
+            ]
+        }
+    },
+    {
+        "id": 44941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001251.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001251.wav"
+            ]
+        }
+    },
+    {
+        "id": 44942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002814.wav"
+            ]
+        }
+    },
+    {
+        "id": 44954,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000223.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000223.wav"
+            ]
+        }
+    },
+    {
+        "id": 44965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000810.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000810.wav"
+            ]
+        }
+    },
+    {
+        "id": 44975,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000980.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000980.wav"
+            ]
+        }
+    },
+    {
+        "id": 44981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000401.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000401.wav"
+            ]
+        }
+    },
+    {
+        "id": 44987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001144.wav"
+            ]
+        }
+    },
+    {
+        "id": 44995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001016.wav"
+            ]
+        }
+    },
+    {
+        "id": 44996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001108.wav"
+            ]
+        }
+    },
+    {
+        "id": 44997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007602.wav"
+            ]
+        }
+    },
+    {
+        "id": 45008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001393.wav"
+            ]
+        }
+    },
+    {
+        "id": 45017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000843.wav"
+            ]
+        }
+    },
+    {
+        "id": 45023,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001249.wav"
+            ]
+        }
+    },
+    {
+        "id": 45026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006370.wav"
+            ]
+        }
+    },
+    {
+        "id": 45032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000683.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000683.wav"
+            ]
+        }
+    },
+    {
+        "id": 45038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000772.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000772.wav"
+            ]
+        }
+    },
+    {
+        "id": 45046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000701.wav"
+            ]
+        }
+    },
+    {
+        "id": 45056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000781.wav"
+            ]
+        }
+    },
+    {
+        "id": 45064,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001043.wav"
+            ]
+        }
+    },
+    {
+        "id": 45072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001046.wav"
+            ]
+        }
+    },
+    {
+        "id": 45073,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000508.wav"
+            ]
+        }
+    },
+    {
+        "id": 45080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001210.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001210.wav"
+            ]
+        }
+    },
+    {
+        "id": 45084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001407.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001407.wav"
+            ]
+        }
+    },
+    {
+        "id": 45086,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000958.wav"
+            ]
+        }
+    },
+    {
+        "id": 45088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000400.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000400.wav"
+            ]
+        }
+    },
+    {
+        "id": 45091,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000494.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000494.wav"
+            ]
+        }
+    },
+    {
+        "id": 45095,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002353.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002353.wav"
+            ]
+        }
+    },
+    {
+        "id": 45101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005921.wav"
+            ]
+        }
+    },
+    {
+        "id": 45105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005905.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005905.wav"
+            ]
+        }
+    },
+    {
+        "id": 45113,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000018.wav"
+            ]
+        }
+    },
+    {
+        "id": 45119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000660.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000660.wav"
+            ]
+        }
+    },
+    {
+        "id": 45128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000232.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000232.wav"
+            ]
+        }
+    },
+    {
+        "id": 45129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000146.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000146.wav"
+            ]
+        }
+    },
+    {
+        "id": 45150,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001144.wav"
+            ]
+        }
+    },
+    {
+        "id": 45159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000910.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000910.wav"
+            ]
+        }
+    },
+    {
+        "id": 45163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001248.wav"
+            ]
+        }
+    },
+    {
+        "id": 45167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000201.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000201.wav"
+            ]
+        }
+    },
+    {
+        "id": 45169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001484.wav"
+            ]
+        }
+    },
+    {
+        "id": 45171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000753.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000753.wav"
+            ]
+        }
+    },
+    {
+        "id": 45189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000859.wav"
+            ]
+        }
+    },
+    {
+        "id": 45190,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000297.wav"
+            ]
+        }
+    },
+    {
+        "id": 45197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum on the left more rhythmic than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001248.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001248.wav"
+            ]
+        }
+    },
+    {
+        "id": 45198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000535.wav"
+            ]
+        }
+    },
+    {
+        "id": 45202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000349.wav"
+            ]
+        }
+    },
+    {
+        "id": 45209,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001129.wav"
+            ]
+        }
+    },
+    {
+        "id": 45212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001047.wav"
+            ]
+        }
+    },
+    {
+        "id": 45218,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007321.wav"
+            ]
+        }
+    },
+    {
+        "id": 45234,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000636.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000636.wav"
+            ]
+        }
+    },
+    {
+        "id": 45236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right more rhythmic than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008467.wav"
+            ]
+        }
+    },
+    {
+        "id": 45238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001143.wav"
+            ]
+        }
+    },
+    {
+        "id": 45241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left louder than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001298.wav"
+            ]
+        }
+    },
+    {
+        "id": 45244,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001115.wav"
+            ]
+        }
+    },
+    {
+        "id": 45257,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000791.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000791.wav"
+            ]
+        }
+    },
+    {
+        "id": 45262,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000464.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000464.wav"
+            ]
+        }
+    },
+    {
+        "id": 45266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007205.wav"
+            ]
+        }
+    },
+    {
+        "id": 45267,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000001.wav"
+            ]
+        }
+    },
+    {
+        "id": 45275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001289.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001289.wav"
+            ]
+        }
+    },
+    {
+        "id": 45278,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001468.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001468.wav"
+            ]
+        }
+    },
+    {
+        "id": 45283,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000941.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000941.wav"
+            ]
+        }
+    },
+    {
+        "id": 45284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000868.wav"
+            ]
+        }
+    },
+    {
+        "id": 45287,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001038.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001038.wav"
+            ]
+        }
+    },
+    {
+        "id": 45288,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000226.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000226.wav"
+            ]
+        }
+    },
+    {
+        "id": 45299,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001233.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001233.wav"
+            ]
+        }
+    },
+    {
+        "id": 45301,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000946.wav"
+            ]
+        }
+    },
+    {
+        "id": 45303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000670.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000670.wav"
+            ]
+        }
+    },
+    {
+        "id": 45305,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000204.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000204.wav"
+            ]
+        }
+    },
+    {
+        "id": 45312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000972.wav"
+            ]
+        }
+    },
+    {
+        "id": 45335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001085.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001085.wav"
+            ]
+        }
+    },
+    {
+        "id": 45337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005091.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005091.wav"
+            ]
+        }
+    },
+    {
+        "id": 45341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001070.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001070.wav"
+            ]
+        }
+    },
+    {
+        "id": 45345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001237.wav"
+            ]
+        }
+    },
+    {
+        "id": 45348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001030.wav"
+            ]
+        }
+    },
+    {
+        "id": 45360,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the left more rhythmic than the guzheng on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007518.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007518.wav"
+            ]
+        }
+    },
+    {
+        "id": 45364,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the clarinet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000984.wav"
+            ]
+        }
+    },
+    {
+        "id": 45370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000237.wav"
+            ]
+        }
+    },
+    {
+        "id": 45372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000952.wav"
+            ]
+        }
+    },
+    {
+        "id": 45377,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000599.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000599.wav"
+            ]
+        }
+    },
+    {
+        "id": 45381,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000796.wav"
+            ]
+        }
+    },
+    {
+        "id": 45386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002350.wav"
+            ]
+        }
+    },
+    {
+        "id": 45393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000184.wav"
+            ]
+        }
+    },
+    {
+        "id": 45396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001256.wav"
+            ]
+        }
+    },
+    {
+        "id": 45400,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the middle more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000742.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000742.wav"
+            ]
+        }
+    },
+    {
+        "id": 45426,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000800.wav"
+            ]
+        }
+    },
+    {
+        "id": 45429,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001467.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001467.wav"
+            ]
+        }
+    },
+    {
+        "id": 45440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001420.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001420.wav"
+            ]
+        }
+    },
+    {
+        "id": 45446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000331.wav"
+            ]
+        }
+    },
+    {
+        "id": 45447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001172.wav"
+            ]
+        }
+    },
+    {
+        "id": 45453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001340.wav"
+            ]
+        }
+    },
+    {
+        "id": 45467,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000167.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000167.wav"
+            ]
+        }
+    },
+    {
+        "id": 45470,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001442.wav"
+            ]
+        }
+    },
+    {
+        "id": 45471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001157.wav"
+            ]
+        }
+    },
+    {
+        "id": 45472,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000800.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000800.wav"
+            ]
+        }
+    },
+    {
+        "id": 45475,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000974.wav"
+            ]
+        }
+    },
+    {
+        "id": 45478,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000626.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000626.wav"
+            ]
+        }
+    },
+    {
+        "id": 45479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000852.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000852.wav"
+            ]
+        }
+    },
+    {
+        "id": 45480,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007713.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007713.wav"
+            ]
+        }
+    },
+    {
+        "id": 45492,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001172.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001172.wav"
+            ]
+        }
+    },
+    {
+        "id": 45494,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001244.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001244.wav"
+            ]
+        }
+    },
+    {
+        "id": 45501,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000855.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000855.wav"
+            ]
+        }
+    },
+    {
+        "id": 45510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the middle louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000703.wav"
+            ]
+        }
+    },
+    {
+        "id": 45514,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000982.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000982.wav"
+            ]
+        }
+    },
+    {
+        "id": 45522,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the right louder than the clarinet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000818.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000818.wav"
+            ]
+        }
+    },
+    {
+        "id": 45526,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000689.wav"
+            ]
+        }
+    },
+    {
+        "id": 45543,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000360.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000360.wav"
+            ]
+        }
+    },
+    {
+        "id": 45544,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000569.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000569.wav"
+            ]
+        }
+    },
+    {
+        "id": 45548,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001303.wav"
+            ]
+        }
+    },
+    {
+        "id": 45554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the guzheng on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000996.wav"
+            ]
+        }
+    },
+    {
+        "id": 45562,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007205.wav"
+            ]
+        }
+    },
+    {
+        "id": 45567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000540.wav"
+            ]
+        }
+    },
+    {
+        "id": 45576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000826.wav"
+            ]
+        }
+    },
+    {
+        "id": 45586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001047.wav"
+            ]
+        }
+    },
+    {
+        "id": 45601,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008047.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008047.wav"
+            ]
+        }
+    },
+    {
+        "id": 45606,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo louder than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007079.wav"
+            ]
+        }
+    },
+    {
+        "id": 45611,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the suona?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000545.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000545.wav"
+            ]
+        }
+    },
+    {
+        "id": 45617,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001298.wav"
+            ]
+        }
+    },
+    {
+        "id": 45619,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000230.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000230.wav"
+            ]
+        }
+    },
+    {
+        "id": 45623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001132.wav"
+            ]
+        }
+    },
+    {
+        "id": 45624,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002350.wav"
+            ]
+        }
+    },
+    {
+        "id": 45631,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000742.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000742.wav"
+            ]
+        }
+    },
+    {
+        "id": 45633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the banjo on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001477.wav"
+            ]
+        }
+    },
+    {
+        "id": 45634,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000321.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000321.wav"
+            ]
+        }
+    },
+    {
+        "id": 45637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000330.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000330.wav"
+            ]
+        }
+    },
+    {
+        "id": 45652,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000962.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000962.wav"
+            ]
+        }
+    },
+    {
+        "id": 45654,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001252.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001252.wav"
+            ]
+        }
+    },
+    {
+        "id": 45658,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000258.wav"
+            ]
+        }
+    },
+    {
+        "id": 45670,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000581.wav"
+            ]
+        }
+    },
+    {
+        "id": 45671,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001089.wav"
+            ]
+        }
+    },
+    {
+        "id": 45679,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the tuba on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000812.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000812.wav"
+            ]
+        }
+    },
+    {
+        "id": 45691,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001100.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001100.wav"
+            ]
+        }
+    },
+    {
+        "id": 45697,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right more rhythmic than the saxophone on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000587.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000587.wav"
+            ]
+        }
+    },
+    {
+        "id": 45704,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001425.wav"
+            ]
+        }
+    },
+    {
+        "id": 45709,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007357.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007357.wav"
+            ]
+        }
+    },
+    {
+        "id": 45714,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001089.wav"
+            ]
+        }
+    },
+    {
+        "id": 45728,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000272.wav"
+            ]
+        }
+    },
+    {
+        "id": 45738,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006317.wav"
+            ]
+        }
+    },
+    {
+        "id": 45744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001205.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001205.wav"
+            ]
+        }
+    },
+    {
+        "id": 45749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001391.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001391.wav"
+            ]
+        }
+    },
+    {
+        "id": 45750,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000917.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000917.wav"
+            ]
+        }
+    },
+    {
+        "id": 45757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001273.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001273.wav"
+            ]
+        }
+    },
+    {
+        "id": 45768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the left louder than the bagpipe on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000265.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000265.wav"
+            ]
+        }
+    },
+    {
+        "id": 45775,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000217.wav"
+            ]
+        }
+    },
+    {
+        "id": 45782,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000770.wav"
+            ]
+        }
+    },
+    {
+        "id": 45783,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa on the middle louder than the acoustic_guitar on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001197.wav"
+            ]
+        }
+    },
+    {
+        "id": 45806,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000765.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000765.wav"
+            ]
+        }
+    },
+    {
+        "id": 45819,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001310.wav"
+            ]
+        }
+    },
+    {
+        "id": 45822,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000786.wav"
+            ]
+        }
+    },
+    {
+        "id": 45825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001988.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001988.wav"
+            ]
+        }
+    },
+    {
+        "id": 45826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the right louder than the cello on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000845.wav"
+            ]
+        }
+    },
+    {
+        "id": 45834,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000976.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000976.wav"
+            ]
+        }
+    },
+    {
+        "id": 45843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001455.wav"
+            ]
+        }
+    },
+    {
+        "id": 45852,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000458.wav"
+            ]
+        }
+    },
+    {
+        "id": 45861,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001331.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001331.wav"
+            ]
+        }
+    },
+    {
+        "id": 45864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the accordion on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001115.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001115.wav"
+            ]
+        }
+    },
+    {
+        "id": 45866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001381.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001381.wav"
+            ]
+        }
+    },
+    {
+        "id": 45877,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001194.wav"
+            ]
+        }
+    },
+    {
+        "id": 45886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000668.wav"
+            ]
+        }
+    },
+    {
+        "id": 45890,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001448.wav"
+            ]
+        }
+    },
+    {
+        "id": 45892,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005515.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005515.wav"
+            ]
+        }
+    },
+    {
+        "id": 45898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002442.wav"
+            ]
+        }
+    },
+    {
+        "id": 45904,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004789.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004789.wav"
+            ]
+        }
+    },
+    {
+        "id": 45937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000164.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000164.wav"
+            ]
+        }
+    },
+    {
+        "id": 45959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the erhu on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001030.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001030.wav"
+            ]
+        }
+    },
+    {
+        "id": 45967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000938.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000938.wav"
+            ]
+        }
+    },
+    {
+        "id": 45972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000343.wav"
+            ]
+        }
+    },
+    {
+        "id": 45996,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000419.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000419.wav"
+            ]
+        }
+    },
+    {
+        "id": 45997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001005.wav"
+            ]
+        }
+    },
+    {
+        "id": 46003,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000202.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000202.wav"
+            ]
+        }
+    },
+    {
+        "id": 46005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001270.wav"
+            ]
+        }
+    },
+    {
+        "id": 46010,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the right louder than the trumpet on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001455.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001455.wav"
+            ]
+        }
+    },
+    {
+        "id": 46021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000748.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000748.wav"
+            ]
+        }
+    },
+    {
+        "id": 46040,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng on the middle louder than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000936.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000936.wav"
+            ]
+        }
+    },
+    {
+        "id": 46049,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006328.wav"
+            ]
+        }
+    },
+    {
+        "id": 46066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001173.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001173.wav"
+            ]
+        }
+    },
+    {
+        "id": 46069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000906.wav"
+            ]
+        }
+    },
+    {
+        "id": 46072,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000466.wav"
+            ]
+        }
+    },
+    {
+        "id": 46085,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001064.wav"
+            ]
+        }
+    },
+    {
+        "id": 46092,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000963.wav"
+            ]
+        }
+    },
+    {
+        "id": 46096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001387.wav"
+            ]
+        }
+    },
+    {
+        "id": 46097,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000931.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000931.wav"
+            ]
+        }
+    },
+    {
+        "id": 46098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001206.wav"
+            ]
+        }
+    },
+    {
+        "id": 46099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001076.wav"
+            ]
+        }
+    },
+    {
+        "id": 46100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba on the right more rhythmic than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001319.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001319.wav"
+            ]
+        }
+    },
+    {
+        "id": 46105,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000814.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000814.wav"
+            ]
+        }
+    },
+    {
+        "id": 46112,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001108.wav"
+            ]
+        }
+    },
+    {
+        "id": 46121,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001392.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001392.wav"
+            ]
+        }
+    },
+    {
+        "id": 46124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001043.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001043.wav"
+            ]
+        }
+    },
+    {
+        "id": 46130,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005084.wav"
+            ]
+        }
+    },
+    {
+        "id": 46134,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001367.wav"
+            ]
+        }
+    },
+    {
+        "id": 46142,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the middle louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001350.wav"
+            ]
+        }
+    },
+    {
+        "id": 46159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele more rhythmic than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001374.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001374.wav"
+            ]
+        }
+    },
+    {
+        "id": 46163,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000796.wav"
+            ]
+        }
+    },
+    {
+        "id": 46171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000458.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000458.wav"
+            ]
+        }
+    },
+    {
+        "id": 46175,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001361.wav"
+            ]
+        }
+    },
+    {
+        "id": 46181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007457.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007457.wav"
+            ]
+        }
+    },
+    {
+        "id": 46196,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000615.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000615.wav"
+            ]
+        }
+    },
+    {
+        "id": 46197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005623.wav"
+            ]
+        }
+    },
+    {
+        "id": 46202,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000845.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000845.wav"
+            ]
+        }
+    },
+    {
+        "id": 46208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left louder than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001281.wav"
+            ]
+        }
+    },
+    {
+        "id": 46212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001315.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001315.wav"
+            ]
+        }
+    },
+    {
+        "id": 46222,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000723.wav"
+            ]
+        }
+    },
+    {
+        "id": 46223,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006220.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006220.wav"
+            ]
+        }
+    },
+    {
+        "id": 46224,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000151.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000151.wav"
+            ]
+        }
+    },
+    {
+        "id": 46236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001474.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001474.wav"
+            ]
+        }
+    },
+    {
+        "id": 46237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001361.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001361.wav"
+            ]
+        }
+    },
+    {
+        "id": 46241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001476.wav"
+            ]
+        }
+    },
+    {
+        "id": 46246,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003950.wav"
+            ]
+        }
+    },
+    {
+        "id": 46247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002274.wav"
+            ]
+        }
+    },
+    {
+        "id": 46250,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele on the middle more rhythmic than the tuba on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001334.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001334.wav"
+            ]
+        }
+    },
+    {
+        "id": 46251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000576.wav"
+            ]
+        }
+    },
+    {
+        "id": 46252,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005551.wav"
+            ]
+        }
+    },
+    {
+        "id": 46255,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000891.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000891.wav"
+            ]
+        }
+    },
+    {
+        "id": 46260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005322.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005322.wav"
+            ]
+        }
+    },
+    {
+        "id": 46265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000701.wav"
+            ]
+        }
+    },
+    {
+        "id": 46275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the congas playing longer than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000562.wav"
+            ]
+        }
+    },
+    {
+        "id": 46304,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001306.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001306.wav"
+            ]
+        }
+    },
+    {
+        "id": 46307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute louder than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000720.wav"
+            ]
+        }
+    },
+    {
+        "id": 46326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich accordion makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004320.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004320.wav"
+            ]
+        }
+    },
+    {
+        "id": 46327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001009.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001009.wav"
+            ]
+        }
+    },
+    {
+        "id": 46334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001020.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001020.wav"
+            ]
+        }
+    },
+    {
+        "id": 46337,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001076.wav"
+            ]
+        }
+    },
+    {
+        "id": 46339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001189.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001189.wav"
+            ]
+        }
+    },
+    {
+        "id": 46341,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001166.wav"
+            ]
+        }
+    },
+    {
+        "id": 46343,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000952.wav"
+            ]
+        }
+    },
+    {
+        "id": 46349,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the lowest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000622.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000622.wav"
+            ]
+        }
+    },
+    {
+        "id": 46350,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001012.wav"
+            ]
+        }
+    },
+    {
+        "id": 46352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001433.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001433.wav"
+            ]
+        }
+    },
+    {
+        "id": 46357,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000839.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000839.wav"
+            ]
+        }
+    },
+    {
+        "id": 46373,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001029.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001029.wav"
+            ]
+        }
+    },
+    {
+        "id": 46374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000289.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000289.wav"
+            ]
+        }
+    },
+    {
+        "id": 46392,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000287.wav"
+            ]
+        }
+    },
+    {
+        "id": 46393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich cello makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000732.wav"
+            ]
+        }
+    },
+    {
+        "id": 46402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000324.wav"
+            ]
+        }
+    },
+    {
+        "id": 46405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001336.wav"
+            ]
+        }
+    },
+    {
+        "id": 46406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001484.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001484.wav"
+            ]
+        }
+    },
+    {
+        "id": 46438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000699.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000699.wav"
+            ]
+        }
+    },
+    {
+        "id": 46445,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001116.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001116.wav"
+            ]
+        }
+    },
+    {
+        "id": 46446,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001397.wav"
+            ]
+        }
+    },
+    {
+        "id": 46453,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000811.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000811.wav"
+            ]
+        }
+    },
+    {
+        "id": 46457,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000427.wav"
+            ]
+        }
+    },
+    {
+        "id": 46466,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000371.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000371.wav"
+            ]
+        }
+    },
+    {
+        "id": 46476,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001104.wav"
+            ]
+        }
+    },
+    {
+        "id": 46477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the right more rhythmic than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001433.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001433.wav"
+            ]
+        }
+    },
+    {
+        "id": 46483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008079.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008079.wav"
+            ]
+        }
+    },
+    {
+        "id": 46487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001283.wav"
+            ]
+        }
+    },
+    {
+        "id": 46489,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001297.wav"
+            ]
+        }
+    },
+    {
+        "id": 46495,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001336.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001336.wav"
+            ]
+        }
+    },
+    {
+        "id": 46496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000923.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000923.wav"
+            ]
+        }
+    },
+    {
+        "id": 46497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001028.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001028.wav"
+            ]
+        }
+    },
+    {
+        "id": 46500,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000795.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000795.wav"
+            ]
+        }
+    },
+    {
+        "id": 46506,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000416.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000416.wav"
+            ]
+        }
+    },
+    {
+        "id": 46510,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001218.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001218.wav"
+            ]
+        }
+    },
+    {
+        "id": 46520,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the saxophone on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003950.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003950.wav"
+            ]
+        }
+    },
+    {
+        "id": 46523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001237.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001237.wav"
+            ]
+        }
+    },
+    {
+        "id": 46528,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000796.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000796.wav"
+            ]
+        }
+    },
+    {
+        "id": 46531,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006110.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006110.wav"
+            ]
+        }
+    },
+    {
+        "id": 46537,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001056.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001056.wav"
+            ]
+        }
+    },
+    {
+        "id": 46542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005242.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005242.wav"
+            ]
+        }
+    },
+    {
+        "id": 46549,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001217.wav"
+            ]
+        }
+    },
+    {
+        "id": 46552,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000831.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000831.wav"
+            ]
+        }
+    },
+    {
+        "id": 46554,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000725.wav"
+            ]
+        }
+    },
+    {
+        "id": 46557,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007297.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007297.wav"
+            ]
+        }
+    },
+    {
+        "id": 46560,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000397.wav"
+            ]
+        }
+    },
+    {
+        "id": 46561,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000551.wav"
+            ]
+        }
+    },
+    {
+        "id": 46566,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001099.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001099.wav"
+            ]
+        }
+    },
+    {
+        "id": 46567,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000602.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000602.wav"
+            ]
+        }
+    },
+    {
+        "id": 46579,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone louder than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001184.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001184.wav"
+            ]
+        }
+    },
+    {
+        "id": 46586,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000768.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000768.wav"
+            ]
+        }
+    },
+    {
+        "id": 46590,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar louder than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000882.wav"
+            ]
+        }
+    },
+    {
+        "id": 46595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001583.wav"
+            ]
+        }
+    },
+    {
+        "id": 46597,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000930.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000930.wav"
+            ]
+        }
+    },
+    {
+        "id": 46599,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the saxophone on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000947.wav"
+            ]
+        }
+    },
+    {
+        "id": 46612,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008174.wav"
+            ]
+        }
+    },
+    {
+        "id": 46621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello louder than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001097.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001097.wav"
+            ]
+        }
+    },
+    {
+        "id": 46623,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001325.wav"
+            ]
+        }
+    },
+    {
+        "id": 46626,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000781.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000781.wav"
+            ]
+        }
+    },
+    {
+        "id": 46628,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the middle louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000867.wav"
+            ]
+        }
+    },
+    {
+        "id": 46647,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001064.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001064.wav"
+            ]
+        }
+    },
+    {
+        "id": 46649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu louder than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001453.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001453.wav"
+            ]
+        }
+    },
+    {
+        "id": 46650,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000972.wav"
+            ]
+        }
+    },
+    {
+        "id": 46657,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000883.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000883.wav"
+            ]
+        }
+    },
+    {
+        "id": 46663,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008526.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008526.wav"
+            ]
+        }
+    },
+    {
+        "id": 46669,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000329.wav"
+            ]
+        }
+    },
+    {
+        "id": 46673,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the right louder than the erhu on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000310.wav"
+            ]
+        }
+    },
+    {
+        "id": 46686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000193.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000193.wav"
+            ]
+        }
+    },
+    {
+        "id": 46692,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000427.wav"
+            ]
+        }
+    },
+    {
+        "id": 46696,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001238.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001238.wav"
+            ]
+        }
+    },
+    {
+        "id": 46718,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000882.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000882.wav"
+            ]
+        }
+    },
+    {
+        "id": 46737,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000740.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000740.wav"
+            ]
+        }
+    },
+    {
+        "id": 46741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000520.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000520.wav"
+            ]
+        }
+    },
+    {
+        "id": 46744,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008603.wav"
+            ]
+        }
+    },
+    {
+        "id": 46749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001055.wav"
+            ]
+        }
+    },
+    {
+        "id": 46754,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001159.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001159.wav"
+            ]
+        }
+    },
+    {
+        "id": 46756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000806.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000806.wav"
+            ]
+        }
+    },
+    {
+        "id": 46757,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon louder than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001325.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001325.wav"
+            ]
+        }
+    },
+    {
+        "id": 46759,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001312.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001312.wav"
+            ]
+        }
+    },
+    {
+        "id": 46777,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right louder than the piano on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001397.wav"
+            ]
+        }
+    },
+    {
+        "id": 46792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000485.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000485.wav"
+            ]
+        }
+    },
+    {
+        "id": 46797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006510.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006510.wav"
+            ]
+        }
+    },
+    {
+        "id": 46813,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000723.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000723.wav"
+            ]
+        }
+    },
+    {
+        "id": 46815,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008005.wav"
+            ]
+        }
+    },
+    {
+        "id": 46824,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the middle louder than the pipa on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001415.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001415.wav"
+            ]
+        }
+    },
+    {
+        "id": 46825,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000551.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000551.wav"
+            ]
+        }
+    },
+    {
+        "id": 46827,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano louder than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001445.wav"
+            ]
+        }
+    },
+    {
+        "id": 46843,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich pipa makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000827.wav"
+            ]
+        }
+    },
+    {
+        "id": 46847,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002275.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002275.wav"
+            ]
+        }
+    },
+    {
+        "id": 46853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005016.wav"
+            ]
+        }
+    },
+    {
+        "id": 46868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000736.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000736.wav"
+            ]
+        }
+    },
+    {
+        "id": 46869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000668.wav"
+            ]
+        }
+    },
+    {
+        "id": 46871,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum louder than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000425.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000425.wav"
+            ]
+        }
+    },
+    {
+        "id": 46876,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the middle louder than the suona on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001448.wav"
+            ]
+        }
+    },
+    {
+        "id": 46881,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001126.wav"
+            ]
+        }
+    },
+    {
+        "id": 46882,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000933.wav"
+            ]
+        }
+    },
+    {
+        "id": 46887,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001402.wav"
+            ]
+        }
+    },
+    {
+        "id": 46902,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001113.wav"
+            ]
+        }
+    },
+    {
+        "id": 46903,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle louder than the suona on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001341.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001341.wav"
+            ]
+        }
+    },
+    {
+        "id": 46915,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000903.wav"
+            ]
+        }
+    },
+    {
+        "id": 46917,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007512.wav"
+            ]
+        }
+    },
+    {
+        "id": 46930,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000270.wav"
+            ]
+        }
+    },
+    {
+        "id": 46931,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000835.wav"
+            ]
+        }
+    },
+    {
+        "id": 46955,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002907.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002907.wav"
+            ]
+        }
+    },
+    {
+        "id": 46958,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000378.wav"
+            ]
+        }
+    },
+    {
+        "id": 46973,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001108.wav"
+            ]
+        }
+    },
+    {
+        "id": 46976,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the piano on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001310.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001310.wav"
+            ]
+        }
+    },
+    {
+        "id": 46984,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001427.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001427.wav"
+            ]
+        }
+    },
+    {
+        "id": 46986,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000505.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000505.wav"
+            ]
+        }
+    },
+    {
+        "id": 46993,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000945.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000945.wav"
+            ]
+        }
+    },
+    {
+        "id": 47008,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the right louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000835.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000835.wav"
+            ]
+        }
+    },
+    {
+        "id": 47009,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001168.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001168.wav"
+            ]
+        }
+    },
+    {
+        "id": 47012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001076.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001076.wav"
+            ]
+        }
+    },
+    {
+        "id": 47017,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000477.wav"
+            ]
+        }
+    },
+    {
+        "id": 47019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001350.wav"
+            ]
+        }
+    },
+    {
+        "id": 47036,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001126.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001126.wav"
+            ]
+        }
+    },
+    {
+        "id": 47038,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001476.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001476.wav"
+            ]
+        }
+    },
+    {
+        "id": 47043,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000350.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000350.wav"
+            ]
+        }
+    },
+    {
+        "id": 47044,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001098.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001098.wav"
+            ]
+        }
+    },
+    {
+        "id": 47046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000751.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000751.wav"
+            ]
+        }
+    },
+    {
+        "id": 47051,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001083.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001083.wav"
+            ]
+        }
+    },
+    {
+        "id": 47055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001010.wav"
+            ]
+        }
+    },
+    {
+        "id": 47057,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000777.wav"
+            ]
+        }
+    },
+    {
+        "id": 47069,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000678.wav"
+            ]
+        }
+    },
+    {
+        "id": 47090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008454.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008454.wav"
+            ]
+        }
+    },
+    {
+        "id": 47104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000465.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000465.wav"
+            ]
+        }
+    },
+    {
+        "id": 47107,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the middle louder than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001187.wav"
+            ]
+        }
+    },
+    {
+        "id": 47123,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000142.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000142.wav"
+            ]
+        }
+    },
+    {
+        "id": 47124,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000871.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000871.wav"
+            ]
+        }
+    },
+    {
+        "id": 47128,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000946.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000946.wav"
+            ]
+        }
+    },
+    {
+        "id": 47148,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000783.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000783.wav"
+            ]
+        }
+    },
+    {
+        "id": 47149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000215.wav"
+            ]
+        }
+    },
+    {
+        "id": 47155,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001148.wav"
+            ]
+        }
+    },
+    {
+        "id": 47159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000715.wav"
+            ]
+        }
+    },
+    {
+        "id": 47174,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000199.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000199.wav"
+            ]
+        }
+    },
+    {
+        "id": 47181,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001390.wav"
+            ]
+        }
+    },
+    {
+        "id": 47189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001356.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001356.wav"
+            ]
+        }
+    },
+    {
+        "id": 47197,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005678.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005678.wav"
+            ]
+        }
+    },
+    {
+        "id": 47199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000996.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000996.wav"
+            ]
+        }
+    },
+    {
+        "id": 47201,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right more rhythmic than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001067.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001067.wav"
+            ]
+        }
+    },
+    {
+        "id": 47212,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000363.wav"
+            ]
+        }
+    },
+    {
+        "id": 47231,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000715.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000715.wav"
+            ]
+        }
+    },
+    {
+        "id": 47236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich banjo makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008380.wav"
+            ]
+        }
+    },
+    {
+        "id": 47237,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000276.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000276.wav"
+            ]
+        }
+    },
+    {
+        "id": 47238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006658.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006658.wav"
+            ]
+        }
+    },
+    {
+        "id": 47247,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left louder than the cello on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000178.wav"
+            ]
+        }
+    },
+    {
+        "id": 47260,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000462.wav"
+            ]
+        }
+    },
+    {
+        "id": 47270,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001281.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001281.wav"
+            ]
+        }
+    },
+    {
+        "id": 47274,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000249.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000249.wav"
+            ]
+        }
+    },
+    {
+        "id": 47276,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000163.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000163.wav"
+            ]
+        }
+    },
+    {
+        "id": 47281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000695.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000695.wav"
+            ]
+        }
+    },
+    {
+        "id": 47284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000144.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000144.wav"
+            ]
+        }
+    },
+    {
+        "id": 47285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000266.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000266.wav"
+            ]
+        }
+    },
+    {
+        "id": 47293,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001045.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001045.wav"
+            ]
+        }
+    },
+    {
+        "id": 47303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum playing longer than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001187.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001187.wav"
+            ]
+        }
+    },
+    {
+        "id": 47307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass on the left louder than the flute on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000616.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000616.wav"
+            ]
+        }
+    },
+    {
+        "id": 47310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000756.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000756.wav"
+            ]
+        }
+    },
+    {
+        "id": 47326,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000934.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000934.wav"
+            ]
+        }
+    },
+    {
+        "id": 47327,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich banjo makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000815.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000815.wav"
+            ]
+        }
+    },
+    {
+        "id": 47334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000688.wav"
+            ]
+        }
+    },
+    {
+        "id": 47335,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001402.wav"
+            ]
+        }
+    },
+    {
+        "id": 47358,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001097.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001097.wav"
+            ]
+        }
+    },
+    {
+        "id": 47370,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000337.wav"
+            ]
+        }
+    },
+    {
+        "id": 47384,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000463.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000463.wav"
+            ]
+        }
+    },
+    {
+        "id": 47386,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005285.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005285.wav"
+            ]
+        }
+    },
+    {
+        "id": 47387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000669.wav"
+            ]
+        }
+    },
+    {
+        "id": 47393,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001414.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001414.wav"
+            ]
+        }
+    },
+    {
+        "id": 47395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000560.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000560.wav"
+            ]
+        }
+    },
+    {
+        "id": 47396,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right louder than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000349.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000349.wav"
+            ]
+        }
+    },
+    {
+        "id": 47406,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000623.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000623.wav"
+            ]
+        }
+    },
+    {
+        "id": 47410,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000439.wav"
+            ]
+        }
+    },
+    {
+        "id": 47413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000963.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000963.wav"
+            ]
+        }
+    },
+    {
+        "id": 47416,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001121.wav"
+            ]
+        }
+    },
+    {
+        "id": 47422,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001157.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001157.wav"
+            ]
+        }
+    },
+    {
+        "id": 47428,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound last?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001366.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001366.wav"
+            ]
+        }
+    },
+    {
+        "id": 47436,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000535.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000535.wav"
+            ]
+        }
+    },
+    {
+        "id": 47450,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000155.wav"
+            ]
+        }
+    },
+    {
+        "id": 47456,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000296.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000296.wav"
+            ]
+        }
+    },
+    {
+        "id": 47463,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000706.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000706.wav"
+            ]
+        }
+    },
+    {
+        "id": 47464,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000006.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000006.wav"
+            ]
+        }
+    },
+    {
+        "id": 47477,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001261.wav"
+            ]
+        }
+    },
+    {
+        "id": 47479,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000246.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000246.wav"
+            ]
+        }
+    },
+    {
+        "id": 47481,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000610.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000610.wav"
+            ]
+        }
+    },
+    {
+        "id": 47483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000293.wav"
+            ]
+        }
+    },
+    {
+        "id": 47486,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the right louder than the ukulele on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000655.wav"
+            ]
+        }
+    },
+    {
+        "id": 47487,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000178.wav"
+            ]
+        }
+    },
+    {
+        "id": 47496,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet on the left more rhythmic than the bassoon on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001147.wav"
+            ]
+        }
+    },
+    {
+        "id": 47497,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin louder than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007181.wav"
+            ]
+        }
+    },
+    {
+        "id": 47507,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000288.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000288.wav"
+            ]
+        }
+    },
+    {
+        "id": 47509,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000642.wav"
+            ]
+        }
+    },
+    {
+        "id": 47524,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000868.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000868.wav"
+            ]
+        }
+    },
+    {
+        "id": 47529,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left more rhythmic than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000217.wav"
+            ]
+        }
+    },
+    {
+        "id": 47530,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002105.wav"
+            ]
+        }
+    },
+    {
+        "id": 47535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the right instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000278.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000278.wav"
+            ]
+        }
+    },
+    {
+        "id": 47539,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000984.wav"
+            ]
+        }
+    },
+    {
+        "id": 47546,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000984.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000984.wav"
+            ]
+        }
+    },
+    {
+        "id": 47547,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000763.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000763.wav"
+            ]
+        }
+    },
+    {
+        "id": 47551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000340.wav"
+            ]
+        }
+    },
+    {
+        "id": 47553,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet louder than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000707.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000707.wav"
+            ]
+        }
+    },
+    {
+        "id": 47555,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000522.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000522.wav"
+            ]
+        }
+    },
+    {
+        "id": 47558,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001329.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001329.wav"
+            ]
+        }
+    },
+    {
+        "id": 47572,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007055.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007055.wav"
+            ]
+        }
+    },
+    {
+        "id": 47574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004309.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004309.wav"
+            ]
+        }
+    },
+    {
+        "id": 47584,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008256.wav"
+            ]
+        }
+    },
+    {
+        "id": 47592,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006217.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006217.wav"
+            ]
+        }
+    },
+    {
+        "id": 47595,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000367.wav"
+            ]
+        }
+    },
+    {
+        "id": 47596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the left instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000994.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000994.wav"
+            ]
+        }
+    },
+    {
+        "id": 47616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhere is the loudest instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003671.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003671.wav"
+            ]
+        }
+    },
+    {
+        "id": 47618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007576.wav"
+            ]
+        }
+    },
+    {
+        "id": 47619,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the first sound coming from the middle instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007687.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007687.wav"
+            ]
+        }
+    },
+    {
+        "id": 47621,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet playing longer than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001174.wav"
+            ]
+        }
+    },
+    {
+        "id": 47629,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007161.wav"
+            ]
+        }
+    },
+    {
+        "id": 47648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich flute makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005209.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005209.wav"
+            ]
+        }
+    },
+    {
+        "id": 47651,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich erhu makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006307.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006307.wav"
+            ]
+        }
+    },
+    {
+        "id": 47663,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich congas makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005002.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005002.wav"
+            ]
+        }
+    },
+    {
+        "id": 47682,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000261.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000261.wav"
+            ]
+        }
+    },
+    {
+        "id": 47683,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007393.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007393.wav"
+            ]
+        }
+    },
+    {
+        "id": 47690,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000402.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000402.wav"
+            ]
+        }
+    },
+    {
+        "id": 47705,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001264.wav"
+            ]
+        }
+    },
+    {
+        "id": 47706,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele in the video always playing?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001175.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001175.wav"
+            ]
+        }
+    },
+    {
+        "id": 47711,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006108.wav"
+            ]
+        }
+    },
+    {
+        "id": 47715,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000983.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000983.wav"
+            ]
+        }
+    },
+    {
+        "id": 47724,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006927.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006927.wav"
+            ]
+        }
+    },
+    {
+        "id": 47739,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs this sound from the instrument in the video?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008473.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008473.wav"
+            ]
+        }
+    },
+    {
+        "id": 47749,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008440.wav"
+            ]
+        }
+    },
+    {
+        "id": 47756,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich congas makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002689.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002689.wav"
+            ]
+        }
+    },
+    {
+        "id": 47761,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003351.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003351.wav"
+            ]
+        }
+    },
+    {
+        "id": 47765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000720.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000720.wav"
+            ]
+        }
+    },
+    {
+        "id": 47768,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000643.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000643.wav"
+            ]
+        }
+    },
+    {
+        "id": 47790,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich piano makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007785.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007785.wav"
+            ]
+        }
+    },
+    {
+        "id": 47800,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion on the left more rhythmic than the trumpet on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002062.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002062.wav"
+            ]
+        }
+    },
+    {
+        "id": 47802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007543.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007543.wav"
+            ]
+        }
+    },
+    {
+        "id": 47831,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000046.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000046.wav"
+            ]
+        }
+    },
+    {
+        "id": 47833,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002944.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002944.wav"
+            ]
+        }
+    },
+    {
+        "id": 47839,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001506.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001506.wav"
+            ]
+        }
+    },
+    {
+        "id": 47842,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003793.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003793.wav"
+            ]
+        }
+    },
+    {
+        "id": 47853,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005661.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005661.wav"
+            ]
+        }
+    },
+    {
+        "id": 47857,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008054.wav"
+            ]
+        }
+    },
+    {
+        "id": 47866,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003445.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003445.wav"
+            ]
+        }
+    },
+    {
+        "id": 47869,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the congas?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004305.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004305.wav"
+            ]
+        }
+    },
+    {
+        "id": 47872,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007216.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007216.wav"
+            ]
+        }
+    },
+    {
+        "id": 47875,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000539.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000539.wav"
+            ]
+        }
+    },
+    {
+        "id": 47889,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004149.wav"
+            ]
+        }
+    },
+    {
+        "id": 47894,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004861.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004861.wav"
+            ]
+        }
+    },
+    {
+        "id": 47918,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007909.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007909.wav"
+            ]
+        }
+    },
+    {
+        "id": 47923,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004513.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004513.wav"
+            ]
+        }
+    },
+    {
+        "id": 47934,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005317.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005317.wav"
+            ]
+        }
+    },
+    {
+        "id": 47936,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007725.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007725.wav"
+            ]
+        }
+    },
+    {
+        "id": 47937,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005596.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005596.wav"
+            ]
+        }
+    },
+    {
+        "id": 47942,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000376.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000376.wav"
+            ]
+        }
+    },
+    {
+        "id": 47944,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006779.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006779.wav"
+            ]
+        }
+    },
+    {
+        "id": 47956,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005466.wav"
+            ]
+        }
+    },
+    {
+        "id": 47959,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004470.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004470.wav"
+            ]
+        }
+    },
+    {
+        "id": 47964,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00005972.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00005972.wav"
+            ]
+        }
+    },
+    {
+        "id": 47967,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006958.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006958.wav"
+            ]
+        }
+    },
+    {
+        "id": 47972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005998.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005998.wav"
+            ]
+        }
+    },
+    {
+        "id": 47979,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001693.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001693.wav"
+            ]
+        }
+    },
+    {
+        "id": 47981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007583.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007583.wav"
+            ]
+        }
+    },
+    {
+        "id": 47983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001557.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001557.wav"
+            ]
+        }
+    },
+    {
+        "id": 47987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich trumpet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004849.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004849.wav"
+            ]
+        }
+    },
+    {
+        "id": 47995,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005848.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005848.wav"
+            ]
+        }
+    },
+    {
+        "id": 48005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008734.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008734.wav"
+            ]
+        }
+    },
+    {
+        "id": 48006,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003152.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003152.wav"
+            ]
+        }
+    },
+    {
+        "id": 48016,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004121.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004121.wav"
+            ]
+        }
+    },
+    {
+        "id": 48021,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001512.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001512.wav"
+            ]
+        }
+    },
+    {
+        "id": 48027,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008653.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008653.wav"
+            ]
+        }
+    },
+    {
+        "id": 48029,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005089.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005089.wav"
+            ]
+        }
+    },
+    {
+        "id": 48035,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000399.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000399.wav"
+            ]
+        }
+    },
+    {
+        "id": 48050,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004431.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004431.wav"
+            ]
+        }
+    },
+    {
+        "id": 48056,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002829.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002829.wav"
+            ]
+        }
+    },
+    {
+        "id": 48058,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001974.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001974.wav"
+            ]
+        }
+    },
+    {
+        "id": 48060,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004489.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004489.wav"
+            ]
+        }
+    },
+    {
+        "id": 48065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004955.wav"
+            ]
+        }
+    },
+    {
+        "id": 48066,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007703.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007703.wav"
+            ]
+        }
+    },
+    {
+        "id": 48074,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000565.wav"
+            ]
+        }
+    },
+    {
+        "id": 48077,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002717.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002717.wav"
+            ]
+        }
+    },
+    {
+        "id": 48078,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007113.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007113.wav"
+            ]
+        }
+    },
+    {
+        "id": 48081,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003323.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003323.wav"
+            ]
+        }
+    },
+    {
+        "id": 48082,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006519.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006519.wav"
+            ]
+        }
+    },
+    {
+        "id": 48083,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000585.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000585.wav"
+            ]
+        }
+    },
+    {
+        "id": 48084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008701.wav"
+            ]
+        }
+    },
+    {
+        "id": 48090,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002068.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002068.wav"
+            ]
+        }
+    },
+    {
+        "id": 48098,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006565.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006565.wav"
+            ]
+        }
+    },
+    {
+        "id": 48099,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona more rhythmic than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006755.wav"
+            ]
+        }
+    },
+    {
+        "id": 48100,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000287.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000287.wav"
+            ]
+        }
+    },
+    {
+        "id": 48109,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008069.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008069.wav"
+            ]
+        }
+    },
+    {
+        "id": 48116,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003369.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003369.wav"
+            ]
+        }
+    },
+    {
+        "id": 48119,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000309.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000309.wav"
+            ]
+        }
+    },
+    {
+        "id": 48120,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001018.wav"
+            ]
+        }
+    },
+    {
+        "id": 48125,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003667.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003667.wav"
+            ]
+        }
+    },
+    {
+        "id": 48129,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008264.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008264.wav"
+            ]
+        }
+    },
+    {
+        "id": 48146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005635.wav"
+            ]
+        }
+    },
+    {
+        "id": 48151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007997.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007997.wav"
+            ]
+        }
+    },
+    {
+        "id": 48156,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000190.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000190.wav"
+            ]
+        }
+    },
+    {
+        "id": 48158,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001194.wav"
+            ]
+        }
+    },
+    {
+        "id": 48159,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002311.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002311.wav"
+            ]
+        }
+    },
+    {
+        "id": 48162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006178.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006178.wav"
+            ]
+        }
+    },
+    {
+        "id": 48169,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the guzheng?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001418.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001418.wav"
+            ]
+        }
+    },
+    {
+        "id": 48198,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003198.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003198.wav"
+            ]
+        }
+    },
+    {
+        "id": 48199,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001148.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001148.wav"
+            ]
+        }
+    },
+    {
+        "id": 48205,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007016.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007016.wav"
+            ]
+        }
+    },
+    {
+        "id": 48208,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000005.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000005.wav"
+            ]
+        }
+    },
+    {
+        "id": 48210,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003955.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003955.wav"
+            ]
+        }
+    },
+    {
+        "id": 48215,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007390.wav"
+            ]
+        }
+    },
+    {
+        "id": 48225,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002370.wav"
+            ]
+        }
+    },
+    {
+        "id": 48228,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000139.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000139.wav"
+            ]
+        }
+    },
+    {
+        "id": 48229,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion more rhythmic than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004303.wav"
+            ]
+        }
+    },
+    {
+        "id": 48236,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00003001.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00003001.wav"
+            ]
+        }
+    },
+    {
+        "id": 48241,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu on the left more rhythmic than the pipa on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000825.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000825.wav"
+            ]
+        }
+    },
+    {
+        "id": 48242,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the middle more rhythmic than the accordion on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008672.wav"
+            ]
+        }
+    },
+    {
+        "id": 48251,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004370.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004370.wav"
+            ]
+        }
+    },
+    {
+        "id": 48271,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006581.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006581.wav"
+            ]
+        }
+    },
+    {
+        "id": 48292,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000250.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000250.wav"
+            ]
+        }
+    },
+    {
+        "id": 48306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000576.wav"
+            ]
+        }
+    },
+    {
+        "id": 48309,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000921.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000921.wav"
+            ]
+        }
+    },
+    {
+        "id": 48311,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004999.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004999.wav"
+            ]
+        }
+    },
+    {
+        "id": 48313,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000082.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000082.wav"
+            ]
+        }
+    },
+    {
+        "id": 48317,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003120.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003120.wav"
+            ]
+        }
+    },
+    {
+        "id": 48334,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000594.wav"
+            ]
+        }
+    },
+    {
+        "id": 48344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003536.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003536.wav"
+            ]
+        }
+    },
+    {
+        "id": 48352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003256.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003256.wav"
+            ]
+        }
+    },
+    {
+        "id": 48355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the second instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004559.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004559.wav"
+            ]
+        }
+    },
+    {
+        "id": 48363,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007270.wav"
+            ]
+        }
+    },
+    {
+        "id": 48372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000577.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000577.wav"
+            ]
+        }
+    },
+    {
+        "id": 48395,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000824.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000824.wav"
+            ]
+        }
+    },
+    {
+        "id": 48402,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000562.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000562.wav"
+            ]
+        }
+    },
+    {
+        "id": 48403,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000243.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000243.wav"
+            ]
+        }
+    },
+    {
+        "id": 48405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006424.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006424.wav"
+            ]
+        }
+    },
+    {
+        "id": 48407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000324.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000324.wav"
+            ]
+        }
+    },
+    {
+        "id": 48433,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001677.wav"
+            ]
+        }
+    },
+    {
+        "id": 48438,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002386.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002386.wav"
+            ]
+        }
+    },
+    {
+        "id": 48440,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo on the right more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001203.wav"
+            ]
+        }
+    },
+    {
+        "id": 48441,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008222.wav"
+            ]
+        }
+    },
+    {
+        "id": 48447,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001952.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001952.wav"
+            ]
+        }
+    },
+    {
+        "id": 48449,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002262.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002262.wav"
+            ]
+        }
+    },
+    {
+        "id": 48455,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the banjo on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008594.wav"
+            ]
+        }
+    },
+    {
+        "id": 48465,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001367.wav"
+            ]
+        }
+    },
+    {
+        "id": 48469,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002866.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002866.wav"
+            ]
+        }
+    },
+    {
+        "id": 48471,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007677.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007677.wav"
+            ]
+        }
+    },
+    {
+        "id": 48483,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000423.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000423.wav"
+            ]
+        }
+    },
+    {
+        "id": 48484,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin playing longer than the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000482.wav"
+            ]
+        }
+    },
+    {
+        "id": 48493,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000731.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000731.wav"
+            ]
+        }
+    },
+    {
+        "id": 48502,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000439.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000439.wav"
+            ]
+        }
+    },
+    {
+        "id": 48513,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003906.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003906.wav"
+            ]
+        }
+    },
+    {
+        "id": 48523,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007340.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007340.wav"
+            ]
+        }
+    },
+    {
+        "id": 48533,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000012.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000012.wav"
+            ]
+        }
+    },
+    {
+        "id": 48534,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001137.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001137.wav"
+            ]
+        }
+    },
+    {
+        "id": 48535,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007777.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007777.wav"
+            ]
+        }
+    },
+    {
+        "id": 48540,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000440.wav"
+            ]
+        }
+    },
+    {
+        "id": 48542,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000209.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000209.wav"
+            ]
+        }
+    },
+    {
+        "id": 48551,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000486.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000486.wav"
+            ]
+        }
+    },
+    {
+        "id": 48564,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002875.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002875.wav"
+            ]
+        }
+    },
+    {
+        "id": 48574,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007108.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007108.wav"
+            ]
+        }
+    },
+    {
+        "id": 48576,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004598.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004598.wav"
+            ]
+        }
+    },
+    {
+        "id": 48588,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005440.wav"
+            ]
+        }
+    },
+    {
+        "id": 48596,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006294.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006294.wav"
+            ]
+        }
+    },
+    {
+        "id": 48610,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001215.wav"
+            ]
+        }
+    },
+    {
+        "id": 48616,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001462.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001462.wav"
+            ]
+        }
+    },
+    {
+        "id": 48618,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002721.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002721.wav"
+            ]
+        }
+    },
+    {
+        "id": 48630,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000672.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000672.wav"
+            ]
+        }
+    },
+    {
+        "id": 48633,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003867.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003867.wav"
+            ]
+        }
+    },
+    {
+        "id": 48635,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000437.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000437.wav"
+            ]
+        }
+    },
+    {
+        "id": 48637,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008827.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008827.wav"
+            ]
+        }
+    },
+    {
+        "id": 48643,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007442.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007442.wav"
+            ]
+        }
+    },
+    {
+        "id": 48648,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002040.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002040.wav"
+            ]
+        }
+    },
+    {
+        "id": 48649,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007397.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007397.wav"
+            ]
+        }
+    },
+    {
+        "id": 48653,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the bassoon?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002874.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002874.wav"
+            ]
+        }
+    },
+    {
+        "id": 48674,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001263.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001263.wav"
+            ]
+        }
+    },
+    {
+        "id": 48676,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008724.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008724.wav"
+            ]
+        }
+    },
+    {
+        "id": 48686,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000194.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000194.wav"
+            ]
+        }
+    },
+    {
+        "id": 48687,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000594.wav"
+            ]
+        }
+    },
+    {
+        "id": 48694,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa playing longer than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004086.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004086.wav"
+            ]
+        }
+    },
+    {
+        "id": 48701,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005573.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005573.wav"
+            ]
+        }
+    },
+    {
+        "id": 48723,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008889.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008889.wav"
+            ]
+        }
+    },
+    {
+        "id": 48725,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003836.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003836.wav"
+            ]
+        }
+    },
+    {
+        "id": 48741,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001095.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001095.wav"
+            ]
+        }
+    },
+    {
+        "id": 48746,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001390.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001390.wav"
+            ]
+        }
+    },
+    {
+        "id": 48748,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000908.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000908.wav"
+            ]
+        }
+    },
+    {
+        "id": 48758,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007755.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007755.wav"
+            ]
+        }
+    },
+    {
+        "id": 48765,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng playing longer than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007759.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007759.wav"
+            ]
+        }
+    },
+    {
+        "id": 48767,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000975.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000975.wav"
+            ]
+        }
+    },
+    {
+        "id": 48769,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005980.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005980.wav"
+            ]
+        }
+    },
+    {
+        "id": 48780,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001203.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001203.wav"
+            ]
+        }
+    },
+    {
+        "id": 48784,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007215.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007215.wav"
+            ]
+        }
+    },
+    {
+        "id": 48792,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003132.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003132.wav"
+            ]
+        }
+    },
+    {
+        "id": 48793,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003471.wav"
+            ]
+        }
+    },
+    {
+        "id": 48794,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001743.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001743.wav"
+            ]
+        }
+    },
+    {
+        "id": 48797,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000174.wav"
+            ]
+        }
+    },
+    {
+        "id": 48801,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion more rhythmic than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000594.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000594.wav"
+            ]
+        }
+    },
+    {
+        "id": 48802,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000642.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000642.wav"
+            ]
+        }
+    },
+    {
+        "id": 48804,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007735.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007735.wav"
+            ]
+        }
+    },
+    {
+        "id": 48814,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003629.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003629.wav"
+            ]
+        }
+    },
+    {
+        "id": 48826,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001274.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001274.wav"
+            ]
+        }
+    },
+    {
+        "id": 48832,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the tuba playing longer than the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006826.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006826.wav"
+            ]
+        }
+    },
+    {
+        "id": 48836,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds before the trumpet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002603.wav"
+            ]
+        }
+    },
+    {
+        "id": 48837,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008694.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008694.wav"
+            ]
+        }
+    },
+    {
+        "id": 48841,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the drum more rhythmic than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000448.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000448.wav"
+            ]
+        }
+    },
+    {
+        "id": 48845,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005949.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005949.wav"
+            ]
+        }
+    },
+    {
+        "id": 48860,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001007.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001007.wav"
+            ]
+        }
+    },
+    {
+        "id": 48864,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003576.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003576.wav"
+            ]
+        }
+    },
+    {
+        "id": 48867,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003625.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003625.wav"
+            ]
+        }
+    },
+    {
+        "id": 48868,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002732.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002732.wav"
+            ]
+        }
+    },
+    {
+        "id": 48870,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008119.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008119.wav"
+            ]
+        }
+    },
+    {
+        "id": 48878,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006020.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006020.wav"
+            ]
+        }
+    },
+    {
+        "id": 48879,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000363.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000363.wav"
+            ]
+        }
+    },
+    {
+        "id": 48886,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001147.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001147.wav"
+            ]
+        }
+    },
+    {
+        "id": 48888,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the left more rhythmic than the piano on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/evv00001272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/evv00001272.wav"
+            ]
+        }
+    },
+    {
+        "id": 48898,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the middle more rhythmic than the acoustic_guitar on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006283.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006283.wav"
+            ]
+        }
+    },
+    {
+        "id": 48901,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007338.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007338.wav"
+            ]
+        }
+    },
+    {
+        "id": 48919,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005272.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005272.wav"
+            ]
+        }
+    },
+    {
+        "id": 48920,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa playing longer than the erhu?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008430.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008430.wav"
+            ]
+        }
+    },
+    {
+        "id": 48925,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006568.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006568.wav"
+            ]
+        }
+    },
+    {
+        "id": 48928,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo playing longer than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004036.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004036.wav"
+            ]
+        }
+    },
+    {
+        "id": 48929,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000933.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000933.wav"
+            ]
+        }
+    },
+    {
+        "id": 48941,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000508.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000508.wav"
+            ]
+        }
+    },
+    {
+        "id": 48945,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000477.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000477.wav"
+            ]
+        }
+    },
+    {
+        "id": 48952,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the flute?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008532.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008532.wav"
+            ]
+        }
+    },
+    {
+        "id": 48962,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002359.wav"
+            ]
+        }
+    },
+    {
+        "id": 48965,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000214.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000214.wav"
+            ]
+        }
+    },
+    {
+        "id": 48970,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000222.wav"
+            ]
+        }
+    },
+    {
+        "id": 48972,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008180.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008180.wav"
+            ]
+        }
+    },
+    {
+        "id": 48974,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001104.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001104.wav"
+            ]
+        }
+    },
+    {
+        "id": 48981,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the acoustic_guitar more rhythmic than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003054.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003054.wav"
+            ]
+        }
+    },
+    {
+        "id": 48983,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005129.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005129.wav"
+            ]
+        }
+    },
+    {
+        "id": 48987,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000280.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000280.wav"
+            ]
+        }
+    },
+    {
+        "id": 48991,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007525.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007525.wav"
+            ]
+        }
+    },
+    {
+        "id": 48997,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004143.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004143.wav"
+            ]
+        }
+    },
+    {
+        "id": 49000,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000885.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000885.wav"
+            ]
+        }
+    },
+    {
+        "id": 49005,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu playing longer than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000652.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000652.wav"
+            ]
+        }
+    },
+    {
+        "id": 49007,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000443.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000443.wav"
+            ]
+        }
+    },
+    {
+        "id": 49012,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe playing longer than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008654.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008654.wav"
+            ]
+        }
+    },
+    {
+        "id": 49013,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich ukulele makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005739.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005739.wav"
+            ]
+        }
+    },
+    {
+        "id": 49019,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bagpipe more rhythmic than the pipa?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004786.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004786.wav"
+            ]
+        }
+    },
+    {
+        "id": 49020,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00007328.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00007328.wav"
+            ]
+        }
+    },
+    {
+        "id": 49025,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000449.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000449.wav"
+            ]
+        }
+    },
+    {
+        "id": 49026,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007947.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007947.wav"
+            ]
+        }
+    },
+    {
+        "id": 49028,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001785.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001785.wav"
+            ]
+        }
+    },
+    {
+        "id": 49032,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone more rhythmic than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004130.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004130.wav"
+            ]
+        }
+    },
+    {
+        "id": 49033,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the banjo?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003015.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003015.wav"
+            ]
+        }
+    },
+    {
+        "id": 49046,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005681.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005681.wav"
+            ]
+        }
+    },
+    {
+        "id": 49055,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002440.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002440.wav"
+            ]
+        }
+    },
+    {
+        "id": 49065,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002764.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002764.wav"
+            ]
+        }
+    },
+    {
+        "id": 49071,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008078.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008078.wav"
+            ]
+        }
+    },
+    {
+        "id": 49075,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the bassoon more rhythmic than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000222.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000222.wav"
+            ]
+        }
+    },
+    {
+        "id": 49076,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001018.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001018.wav"
+            ]
+        }
+    },
+    {
+        "id": 49080,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005213.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005213.wav"
+            ]
+        }
+    },
+    {
+        "id": 49084,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the xylophone more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008387.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008387.wav"
+            ]
+        }
+    },
+    {
+        "id": 49087,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet playing longer than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000010.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000010.wav"
+            ]
+        }
+    },
+    {
+        "id": 49088,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002169.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002169.wav"
+            ]
+        }
+    },
+    {
+        "id": 49096,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich instrument makes sounds after the accordion?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000924.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000924.wav"
+            ]
+        }
+    },
+    {
+        "id": 49101,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006590.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006590.wav"
+            ]
+        }
+    },
+    {
+        "id": 49104,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone on the left more rhythmic than the violin on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001915.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001915.wav"
+            ]
+        }
+    },
+    {
+        "id": 49117,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005333.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005333.wav"
+            ]
+        }
+    },
+    {
+        "id": 49133,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006031.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006031.wav"
+            ]
+        }
+    },
+    {
+        "id": 49137,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007821.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007821.wav"
+            ]
+        }
+    },
+    {
+        "id": 49138,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000859.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000859.wav"
+            ]
+        }
+    },
+    {
+        "id": 49141,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000801.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000801.wav"
+            ]
+        }
+    },
+    {
+        "id": 49144,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000436.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000436.wav"
+            ]
+        }
+    },
+    {
+        "id": 49146,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the left more rhythmic than the bassoon on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008034.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008034.wav"
+            ]
+        }
+    },
+    {
+        "id": 49149,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute playing longer than the violin?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000149.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000149.wav"
+            ]
+        }
+    },
+    {
+        "id": 49151,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the suona on the right more rhythmic than the flute on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001613.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001613.wav"
+            ]
+        }
+    },
+    {
+        "id": 49161,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000174.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000174.wav"
+            ]
+        }
+    },
+    {
+        "id": 49162,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000084.wav"
+            ]
+        }
+    },
+    {
+        "id": 49165,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000970.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000970.wav"
+            ]
+        }
+    },
+    {
+        "id": 49167,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002466.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002466.wav"
+            ]
+        }
+    },
+    {
+        "id": 49168,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00006669.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00006669.wav"
+            ]
+        }
+    },
+    {
+        "id": 49171,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the piano playing longer than the clarinet?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005471.wav"
+            ]
+        }
+    },
+    {
+        "id": 49172,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the clarinet on the right more rhythmic than the piano on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005471.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005471.wav"
+            ]
+        }
+    },
+    {
+        "id": 49179,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the guzheng more rhythmic than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000259.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000259.wav"
+            ]
+        }
+    },
+    {
+        "id": 49180,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhich clarinet makes the sound first?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001211.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001211.wav"
+            ]
+        }
+    },
+    {
+        "id": 49189,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the ukulele playing longer than the acoustic_guitar?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003969.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003969.wav"
+            ]
+        }
+    },
+    {
+        "id": 49204,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004603.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004603.wav"
+            ]
+        }
+    },
+    {
+        "id": 49238,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the third instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008747.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008747.wav"
+            ]
+        }
+    },
+    {
+        "id": 49254,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the last sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001197.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001197.wav"
+            ]
+        }
+    },
+    {
+        "id": 49256,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello on the right more rhythmic than the violin on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000017.wav"
+            ]
+        }
+    },
+    {
+        "id": 49261,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000504.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000504.wav"
+            ]
+        }
+    },
+    {
+        "id": 49262,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00002888.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00002888.wav"
+            ]
+        }
+    },
+    {
+        "id": 49265,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00008534.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00008534.wav"
+            ]
+        }
+    },
+    {
+        "id": 49266,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00000843.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00000843.wav"
+            ]
+        }
+    },
+    {
+        "id": 49268,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003105.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003105.wav"
+            ]
+        }
+    },
+    {
+        "id": 49273,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005770.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005770.wav"
+            ]
+        }
+    },
+    {
+        "id": 49275,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000270.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000270.wav"
+            ]
+        }
+    },
+    {
+        "id": 49281,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000337.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000337.wav"
+            ]
+        }
+    },
+    {
+        "id": 49282,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the middle instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006728.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006728.wav"
+            ]
+        }
+    },
+    {
+        "id": 49284,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003482.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003482.wav"
+            ]
+        }
+    },
+    {
+        "id": 49285,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00004863.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00004863.wav"
+            ]
+        }
+    },
+    {
+        "id": 49290,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008367.wav"
+            ]
+        }
+    },
+    {
+        "id": 49291,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the trumpet more rhythmic than the saxophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008284.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008284.wav"
+            ]
+        }
+    },
+    {
+        "id": 49292,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass more rhythmic than the tuba?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001084.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001084.wav"
+            ]
+        }
+    },
+    {
+        "id": 49294,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the flute more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000258.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000258.wav"
+            ]
+        }
+    },
+    {
+        "id": 49296,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/vv00000158.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/vv00000158.wav"
+            ]
+        }
+    },
+    {
+        "id": 49302,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007438.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007438.wav"
+            ]
+        }
+    },
+    {
+        "id": 49303,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000303.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000303.wav"
+            ]
+        }
+    },
+    {
+        "id": 49306,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the electric_bass more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006956.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006956.wav"
+            ]
+        }
+    },
+    {
+        "id": 49307,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000684.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000684.wav"
+            ]
+        }
+    },
+    {
+        "id": 49310,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007355.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007355.wav"
+            ]
+        }
+    },
+    {
+        "id": 49312,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the banjo more rhythmic than the xylophone?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00005166.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00005166.wav"
+            ]
+        }
+    },
+    {
+        "id": 49314,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000293.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000293.wav"
+            ]
+        }
+    },
+    {
+        "id": 49315,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000655.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000655.wav"
+            ]
+        }
+    },
+    {
+        "id": 49316,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002017.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002017.wav"
+            ]
+        }
+    },
+    {
+        "id": 49318,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000155.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000155.wav"
+            ]
+        }
+    },
+    {
+        "id": 49320,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003359.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003359.wav"
+            ]
+        }
+    },
+    {
+        "id": 49323,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000332.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000332.wav"
+            ]
+        }
+    },
+    {
+        "id": 49324,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000299.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000299.wav"
+            ]
+        }
+    },
+    {
+        "id": 49325,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004656.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004656.wav"
+            ]
+        }
+    },
+    {
+        "id": 49328,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the erhu more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/va00003298.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/va00003298.wav"
+            ]
+        }
+    },
+    {
+        "id": 49330,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001702.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001702.wav"
+            ]
+        }
+    },
+    {
+        "id": 49332,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000540.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000540.wav"
+            ]
+        }
+    },
+    {
+        "id": 49339,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00006088.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00006088.wav"
+            ]
+        }
+    },
+    {
+        "id": 49340,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004042.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004042.wav"
+            ]
+        }
+    },
+    {
+        "id": 49344,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000318.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000318.wav"
+            ]
+        }
+    },
+    {
+        "id": 49345,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000701.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000701.wav"
+            ]
+        }
+    },
+    {
+        "id": 49348,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003834.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003834.wav"
+            ]
+        }
+    },
+    {
+        "id": 49352,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001208.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001208.wav"
+            ]
+        }
+    },
+    {
+        "id": 49353,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00000668.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00000668.wav"
+            ]
+        }
+    },
+    {
+        "id": 49355,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the pipa more rhythmic than the ukulele?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/sa00008206.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/sa00008206.wav"
+            ]
+        }
+    },
+    {
+        "id": 49359,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000688.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000688.wav"
+            ]
+        }
+    },
+    {
+        "id": 49362,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000635.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000635.wav"
+            ]
+        }
+    },
+    {
+        "id": 49365,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00003292.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00003292.wav"
+            ]
+        }
+    },
+    {
+        "id": 49372,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the accordion more rhythmic than the drum?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00004903.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00004903.wav"
+            ]
+        }
+    },
+    {
+        "id": 49374,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001161.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001161.wav"
+            ]
+        }
+    },
+    {
+        "id": 49378,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001344.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001344.wav"
+            ]
+        }
+    },
+    {
+        "id": 49380,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000378.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000378.wav"
+            ]
+        }
+    },
+    {
+        "id": 49383,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the left louder than the instrument on the right?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002880.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002880.wav"
+            ]
+        }
+    },
+    {
+        "id": 49385,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the right instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00002176.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00002176.wav"
+            ]
+        }
+    },
+    {
+        "id": 49387,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin more rhythmic than the piano?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00001343.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00001343.wav"
+            ]
+        }
+    },
+    {
+        "id": 49389,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the first instrument that comes in?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000483.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000483.wav"
+            ]
+        }
+    },
+    {
+        "id": 49405,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nWhat is the left instrument of the first sounding instrument?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00007498.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00007498.wav"
+            ]
+        }
+    },
+    {
+        "id": 49407,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the saxophone playing longer than the cello?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000632.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000632.wav"
+            ]
+        }
+    },
+    {
+        "id": 49413,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the violin on the left more rhythmic than the cello on the middle?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/esa00000380.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/esa00000380.wav"
+            ]
+        }
+    },
+    {
+        "id": 49414,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the cello playing longer than the electric_bass?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/00005367.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/00005367.wav"
+            ]
+        }
+    },
+    {
+        "id": 49431,
+        "conversations": [
+            {
+                "from": "human",
+                "value": "Based on the video <video> and audio <audio>, \nIs the instrument on the right louder than the instrument on the left?\nAnswer the question using a single word."
+            },
+            {
+                "from": "gpt",
+                "value": null
+            }
+        ],
+        "modal_inputs": {
+            "video": [
+                "data/evaluation_datasets/MUSIC-AVQA/video/eva00001181.mp4"
+            ],
+            "audio": [
+                "data/evaluation_datasets/MUSIC-AVQA/audio/eva00001181.wav"
+            ]
+        }
+    }
+]
\ No newline at end of file