metadata
license: cc
language:
- en
- zh
pretty_name: ViCaS
configs:
- config_name: default
data_files:
- split: train
path: annotations/v0.1/*.json
ViCaS: A Dataset for Combining Holistic and Pixel-level Video Understanding using Captions with Grounded Segmentation
For details docs and usage guide, refer to the GitHub repo.