CVML-TueAI/grounding-YT-dataset

该数据集包含了视频文件，并且视频文件被压缩成tar格式。数据集分为三种配置：videos, clips, 和 frames，分别代表不同的数据处理方式。具体的数据集内容描述在README中未提供，因此无法给出详细的中文描述。

The dataset consists of video files compressed in tar format. There are three configurations: videos, clips, and frames, each representing a different way of processing the data. Detailed description of the dataset content is not provided in the README, therefore no specific English description can be given.