worldmodelbench / README.md
nielsr's picture
nielsr HF Staff
Initial dataset card for WorldModelBench
ac058bd verified
|
raw
history blame
380 Bytes
metadata
task_categories:
  - video-text-to-text
language:
  - en

This dataset is for evaluating world modeling capabilities of video generation models, as described in the paper WorldModelBench: Judging Video Generation Models As World Models.