VarunGiridhar3/libero40_libero_10_play

该数据集使用LeRobot创建，是一个机器人学习数据集，专门用于模仿学习或强化学习任务。数据集包含200个episodes，总计104000帧，覆盖10个不同任务。机器人类型为panda，数据以parquet格式存储，视频以mp4格式存储，帧率为10 fps。特征包括动作（7维浮点数组）、状态观测（如末端执行器位置、角度和夹爪状态）、两个图像观测（均为256x256 RGB图像），以及时间戳、帧索引、episode索引、任务索引等元数据。数据集仅包含训练分割，总数据文件大小为100 MB，视频文件大小为200 MB。This dataset was created using LeRobot and is a robotics learning dataset designed for imitation learning or reinforcement learning tasks. It contains 200 episodes, totaling 104,000 frames, and covers 10 different tasks. The robot type is panda, with data stored in parquet format and videos in mp4 format at 10 fps. Features include actions (7-dimensional float array), state observations (such as end-effector positions, angles, and gripper states), two image observations (both 256x256 RGB images), and metadata like timestamps, frame indices, episode indices, and task indices. The dataset includes only a training split, with a total data file size of 100 MB and video file size of 200 MB.