Update Repository Information

LICENSE

@@ -0,0 +1,14 @@
+This dataset includes content under multiple licenses:
+
+- Real images from Nechintosh/ghibli: studio-ghibli-nc-license
+- AI images from:
+  - nitrosocke/Ghibli-Diffusion: CreativeML OpenRAIL-M  
+    Loaded with: torch_dtype=torch.float16
+  - KappaNeuro/studio-ghibli-style: CreativeML OpenRAIL++-M (fine-tuned from SDXL)  
+    Loaded with: torch_dtype=torch.float16, variant="fp16"
+  - Note: While the KappaNeuro repository does not explicitly state a license, it is based on Stability AI's SDXL, which is released under CreativeML Open RAIL++-M. Therefore, it is assumed that this model inherits the same license and non-commercial restrictions.
+
+The combined dataset is provided under the most restrictive condition:  
+**Non-commercial research and educational use only. Commercial use is strictly prohibited.**
+
+For details, please refer to the original model and dataset licenses.

README.md

@@ -1,15 +1,51 @@
+---
+size_categories:
+  - 1K<n<10K
+license: other
+tags:
+- ghibli
+- ai-generated
+- image-classification
+language:
+- en
+pretty_name: Ghibli Real vs AI Dataset
+task_categories:
+- image-classification
+task_ids:
+- multi-class-classification
+splits:
+  - name: train
+    num_examples: 4347
+annotations_creators:
+- machine-generated
+source_datasets:
+  - Nechintosh/ghibli
+  - nitrosocke/Ghibli-Diffusion
+  - KappaNeuro/studio-ghibli-style
+dataset_info:
+  labels:
+  - real
+  - ai
+---
+
 # Ghibli Real vs AI-Generated Dataset
 
 This dataset is provided in two forms:
 
-### 1. `default.jsonl`
+### 1. (default) `metadata.jsonl`
 - One sample per line
-- Includes: `image`, `label`, `description`, `pair_id`
+- Includes: `id`, `image`, `label`, `description`
 - Use this for standard classification or image-text training
 
+- Real images sourced from [Nechintosh/ghibli](https://huggingface.co/datasets/Nechintosh/ghibli) (810 images)
+- AI-generated images created using:
+  - [nitrosocke/Ghibli-Diffusion](https://huggingface.co/nitrosocke/Ghibli-Diffusion) (2727 images)
+  - [KappaNeuro/studio-ghibli-style](https://huggingface.co/KappaNeuro/studio-ghibli-style) (810 images)
+  - Note: While the KappaNeuro repository does not explicitly state a license, it is a fine-tuned model based on [Stable Diffusion XL](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0), which is distributed under the CreativeML Open RAIL++-M License. Therefore, it is assumed that this model inherits the same license and non-commercial restrictions.
+
 ### 2. `pairs.jsonl`
 - Real and fake images paired together
-- Includes: `real_image`, `fake_image`, shared `description`, `pair_id`
+- Includes: `real_image`, `ai_image`, shared `description`, `seed`
 - Use this for contrastive learning or meta-learning (e.g., ProtoNet)
 
 ### How to load
@@ -18,8 +54,40 @@ This dataset is provided in two forms:
 from datasets import load_dataset
 
 # Single image classification
-samples = load_dataset("pulnip/ghibli-dataset", data_files="default.jsonl", split="train")
+samples = load_dataset("pulnip/ghibli-dataset", split="train")
 
 # Paired meta-learning structure
 pairs = load_dataset("pulnip/ghibli-dataset", data_files="pairs.jsonl", split="train")
-```
+
+# Convert labels to binary classification: 'real' vs 'ai'
+# Note: The original "label" field contains "real", "nitrosocke", and "KappaNeuro".
+#       You can treat all non-"real" labels as "ai" to use this dataset for binary classification.
+for sample in samples:
+    sample["binary_label"] = "real" if sample["label"] == "real" else "ai"
+```
+
+## License and Usage
+
+This dataset combines data from multiple sources. Please review the licensing conditions carefully.
+
+### Real Images
+- Source: [Nechintosh/ghibli](https://huggingface.co/datasets/Nechintosh/ghibli)
+- License: Not explicitly stated; assumed for **non-commercial research use only**
+
+### AI-Generated Images
+- Source models:
+  - [nitrosocke/Ghibli-Diffusion](https://huggingface.co/nitrosocke/Ghibli-Diffusion)  
+    Loaded with: `torch_dtype=torch.float16`
+  - [KappaNeuro/studio-ghibli-style](https://huggingface.co/KappaNeuro/studio-ghibli-style)  
+    Loaded with: `torch_dtype=torch.float16, variant="fp16"`
+- These models are provided under community licenses that generally restrict usage to **non-commercial and research purposes**.
+
+---
+
+### Summary
+
+This repository is not published under a single license such as MIT.  
+Because the dataset includes content from multiple sources with varying restrictions,  
+**the dataset is licensed under 'other' and should be treated as non-commercial research-use only.**
+
+Users are responsible for reviewing each component’s license terms before redistribution or adaptation.

scripts/gen_img.py

@@ -36,7 +36,7 @@ NUM_IMAGES = 3
 out_dir = f"data/{ID_PREFIX}"
 os.makedirs(out_dir, exist_ok=True)
 
-with open("default.jsonl", "r", encoding="utf-8") as fin, \
+with open("metadata.jsonl", "r", encoding="utf-8") as fin, \
     open("ai_entries.jsonl", "w", encoding="utf-8") as fout, \
     open("pairs.jsonl", "a", encoding="utf-8") as pairs:
     for i, line in enumerate(fin):

scripts/save_img.py

@@ -18,7 +18,7 @@ samples.sort(key=lambda s: natural_key(os.path.basename(s["image"]["path"])))
 
 os.makedirs("data/real", exist_ok=True)
 
-with open("default.jsonl", "w") as f:
+with open("metadata.jsonl", "w") as f:
     for i, sample in enumerate(samples):
         caption = sample["caption"]
         src_path: str = sample["image"]["path"]

scripts/test.py

@@ -1,32 +0,0 @@
-import json
-import re
-
-def sort_pairs_by_ai_image(input_path, output_path):
-    # '..._<숫자1>_<숫자2>.jpg' 패턴: prefix는 숫자1 전까지 전부
-    pattern = re.compile(r"(.+)_(\d+)_(\d+)\.jpg$")
-    entries = []
-
-    # 1. 읽어서 (prefix, num1, num2, entry)로 저장
-    with open(input_path, 'r', encoding='utf-8') as infile:
-        for line in infile:
-            entry = json.loads(line)
-            ai_img = entry.get('ai_image', '')
-            m = pattern.match(ai_img)
-            if m:
-                prefix, num1, num2 = m.groups()
-                entries.append((prefix, int(num1), int(num2), entry))
-            else:
-                # 패턴 안 맞으면 맨 뒤로
-                entries.append((ai_img, float('inf'), float('inf'), entry))
-
-    # 2. (prefix, num1, num2) 순서로 정렬
-    entries.sort(key=lambda x: (x[0], x[1], x[2]))
-
-    # 3. new_pairs.jsonl에 기록
-    with open(output_path, 'w', encoding='utf-8') as outfile:
-        for _, _, _, entry in entries:
-            outfile.write(json.dumps(entry, ensure_ascii=False) + '\n')
-
-if __name__ == "__main__":
-    sort_pairs_by_ai_image("pairs.jsonl", "new_pairs.jsonl")
-    print("✅ new_pairs.jsonl 생성 완료했어! 😉")