Datasets:

Modalities:
Text
Formats:
json
ArXiv:
Libraries:
Datasets
Dask
License:
ViCaS / README.md
Ali2500's picture
Update README.md
f80c5d1 verified
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

Project | GitHub | Arxiv

For details docs and usage guide, refer to the GitHub repo.