Datasets:
Tasks:
Visual Question Answering
Formats:
imagefolder
Languages:
English
Size:
1K - 10K
ArXiv:
Tags:
vision-language-model
physical-reasoning
video-understanding
physical-intelligence
multimodal
benchmark
License:
Update dataset card: add paper link and video-text-to-text task category
#2
by nielsr HF Staff - opened
This PR updates the PhysVista dataset card by:
- Replacing the placeholder paper URL with the actual paper page: https://huggingface.co/papers/2610.00559
- Adding
video-text-to-textto thetask_categoriesmetadata to better reflect the video-based, multimodal nature of the benchmark.
The rest of the dataset card remains unchanged since it already contains detailed documentation, usage instructions, and references.