Ali2500/ViCaS
收藏官方服务：
资源简介：
ViCaS是一个用于结合整体和像素级视频理解的数据集，使用带有基础分割的标题。
The ViCaS dataset is designed for combining holistic and pixel-level video understanding using captions with grounded segmentation. The dataset supports English and Chinese languages, with a default configuration and training data files located at annotations/v0.1/*.json.
提供机构：
Ali2500
