A single bundle (CogCanvas.zip) containing the reference images and per-scene prompt/annotation files used to build composite "person + object + background" generation prompts.

Top-level layout

CogCanvas_dataset/ ├── background/ ├── celeb/ ├── objects/ └── prompt/

Every file path that appears inside the JSON annotations is relative to the zip root (e.g. celeb/CelebA-HQ-img/13288.jpg, objects/images/airpods_pro_2/img_20.jpg… See the full description on the dataset page: https://huggingface.co/datasets/multimedia-synergy-lab/CogCanvas.