Fix task_categories: remove invalid scene-understanding, add mask-generation

Update README metadata

Update README with proper dataset card metadata and license information

Add Visual Jenga datasets: ClutteredParse, COCO, Full Scene Decom, and NYU with HTML visualizations

Upload clutteredparse/32/A.png with huggingface_hub

Upload clutteredparse/32/B.png with huggingface_hub

Upload clutteredparse/32/img.jpg with huggingface_hub

Upload clutteredparse/04/A.png with huggingface_hub

Upload clutteredparse/04/B.png with huggingface_hub

Upload clutteredparse/04/img.jpg with huggingface_hub

Upload clutteredparse/03/A.png with huggingface_hub

Upload clutteredparse/03/B.png with huggingface_hub

Upload clutteredparse/03/img.jpg with huggingface_hub

Upload dataset_coco.html with huggingface_hub

Upload dataset_nyu.html with huggingface_hub

Upload dataset_clutteredparse.html with huggingface_hub

Upload README.md with huggingface_hub

Upload dataset_full_scene_decom.html with huggingface_hub

initial commit