VarunGiridhar3/libero40_libero_object_v2_play

这是一个用于机器人学习的数据集，通过LeRobot工具创建，专门针对panda机器人。数据集包含10个不同的任务，总共有420个剧集和117,600帧数据。数据以parquet文件格式存储，总数据大小为100MB，视频文件大小为200MB，帧率为10fps。特征包括动作（7维浮点数组）、状态观察（如末端执行器位置、姿态和夹爪状态，共8维浮点数组）、两个图像观察（每个为256x256x3的图像）、时间戳、帧索引、剧集索引、索引和任务索引。数据集仅包含训练分割，覆盖所有剧集。This is a dataset for robotic learning, constructed using the LeRobot tool and specifically tailored for the Panda robot. The dataset encompasses 10 distinct tasks, with a total of 420 episodes and 117,600 frames of data. The data is stored in Parquet file format, with a total data size of 100 MB, while the video files have a size of 200 MB, and the frame rate is set to 10 fps. The dataset features include: actions (7-dimensional floating-point array), state observations (such as end-effector position, orientation, and gripper status, which form an 8-dimensional floating-point array in total), two image observations (each being a 256×256×3 RGB image), timestamps, frame indices, episode indices, global indices, and task indices. The dataset only contains a training split that covers all episodes.