StoryVideoQA: Scaling Deep Video Understanding with a Large-Scale, Multi-Genre and Auto-Generated Dataset
Official data repository for StoryVideoQA, presented in the paper StoryVideoQA: Scaling Deep Video Understanding with a Large-Scale, Multi-Genre and Auto-Generated Dataset, accepted by IJCV 2026.
StoryVideoQA is a large-scale, multi-genre dataset designed for Deep Video Understanding (DVU). It features over 363K question-answer pairs (QAs)… See the full description on the dataset page: https://huggingface.co/datasets/ZQFive/StoryVideoQA.
