amanpandey23/table_bussing_02

这是一个用于机器人学任务的数据集，包含100个剧集，共46295帧。数据集包含单一任务，没有视频文件，只有一个数据块，每块1000条数据。数据集的采样率为30帧每秒，并提供训练数据分割。数据集中的特征包括三种图像（左、中、右），以及状态和动作信息，所有这些特征均以Parquet格式存储。This dataset is for robotics tasks, containing 100 episodes with a total of 46,295 frames. The dataset includes one task, with no video files, one data chunk of 1,000 entries. The dataset is sampled at 30 frames per second and provides a training data split. Features in the dataset include three types of images (left, middle, right), as well as state and action information, all stored in Parquet format.