Update app.py

app.py

@@ -11,21 +11,21 @@ multimodal_phi2 = MultiModalPhi2(
     device="cpu",
 )
 
-def add_content(chatbot, input_data, input_type) -> gr.Chatbot:
+def add_content(chatbot, text, image, audio_upload, audio_mic) -> gr.Chatbot:
     textflag, imageflag, audioflag = False, False, False
-    if input_type == "text":
+    if text not in ["", None]:
         chatbot.append((text, None))
         textflag = True
-    if input_type == "image":
+    if image is not None:
         chatbot.append(((image,), None))
         imageflag = True
-    if input_type == "audio":
+    if audio_mic is not None:
         chatbot.append(((audio_mic,), None))
         audioflag = True
-    # else:
-    #     if audio_upload is not None:
-    #         chatbot.append(((audio_upload,), None))
-    #         audioflag = True
+    else:
+        if audio_upload is not None:
+            chatbot.append(((audio_upload,), None))
+            audioflag = True
     if not any([textflag, imageflag, audioflag]):
         # Raise an error if neither text nor file is provided
         raise gr.Error("Enter a valid text, image or audio")
@@ -61,31 +61,54 @@ def run(history, text, image, audio_upload, audio_mic):
 
 
 with gr.Blocks() as demo:
-    chatbot = gr.Chatbot(
-        [],
-        elem_id="chatbot",
-        bubble_full_width=False,
-        avatar_images=(None, (os.path.join(os.path.dirname(__file__), "avatar.png"))),
-    )
+    gr.Markdown("## MulitModal Phi2 Model Pretraining and Finetuning from Scratch")
 
     with gr.Row():
-        txt = gr.Textbox(
-            scale=4,
-            show_label=False,
-            placeholder="Enter text and press enter",
-            container=False,
-        )
-        img_audio = gr.UploadButton("📁", file_types=["image", "audio"], label="Upload Image or Audio")
-
-    txt_msg = txt.submit(add_content, [chatbot, txt], [chatbot, txt, "text"], queue=False).then(
-        bot, chatbot, chatbot, api_name="bot_response"
+        with gr.Column(scale=4):
+            # Creating a column with a scale of 6
+            with gr.Box():
+                with gr.Row():
+                    # Adding a Textbox with a placeholder "write prompt"
+                    prompt = gr.Textbox(
+                        placeholder="Enter Prompt", lines=2, label="Query", value=None
+                    )
+                # Creating a column with a scale of 2
+                with gr.Row():
+                    # Adding image
+                    image = gr.UploadButton("📁", file_types=["image"], label="Upload Image")
+                # Creating a column with a scale of 2
+                with gr.Row():
+                    # Add audio
+                    audio_upload = gr.UploadButton("📁", file_types=["audio"], label="Upload Audio")
+                    audio_mic = gr.Audio(
+                        source="microphone", type="filepath", format="mp3"
+                    )
+
+        with gr.Column(scale=8):
+            with gr.Box():
+                with gr.Row():
+                    chatbot = gr.Chatbot(
+                        avatar_images=("🧑", "🤖"),
+                        height=550,
+                    )
+                with gr.Row():
+                    # Adding a Button
+                    submit = gr.Button()
+                    clear = gr.Button(value="Clear")
+
+    submit.click(
+        add_content,
+        inputs=[chatbot, prompt, image, audio_upload, audio_mic],
+        outputs=[chatbot],
+    ).success(
+        run,
+        inputs=[chatbot, prompt, image, audio_upload, audio_mic],
+        outputs=[chatbot, prompt, image, audio_upload, audio_mic],
     )
-    img_audio_msg = img_audio.upload(add_input, [chatbot, img_audio], [chatbot, "image"], queue=False).then(
-        bot, chatbot, chatbot
-    )
-
-    # chatbot.like(print_like_dislike, None, None)
 
+    clear.click(
+        clear_data,
+        outputs=[prompt, image, audio_upload, audio_mic, chatbot],
+    )
 
-    
-demo.launch()
+demo.launch()